WO2017193933A1 - 交通事故预警方法及交通事故预警装置 - Google Patents

交通事故预警方法及交通事故预警装置 Download PDF

Info

Publication number
WO2017193933A1
WO2017193933A1 PCT/CN2017/083762 CN2017083762W WO2017193933A1 WO 2017193933 A1 WO2017193933 A1 WO 2017193933A1 CN 2017083762 W CN2017083762 W CN 2017083762W WO 2017193933 A1 WO2017193933 A1 WO 2017193933A1
Authority
WO
WIPO (PCT)
Prior art keywords
traffic
scene
video
traffic accident
target area
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2017/083762
Other languages
English (en)
French (fr)
Inventor
吴一凡
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tencent Technology (Shenzhen) Co Ltd
Original Assignee
Tencent Technology (Shenzhen) Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tencent Technology (Shenzhen) Co Ltd filed Critical Tencent Technology (Shenzhen) Co Ltd
Priority to EP17795551.5A priority Critical patent/EP3457380A4/en
Publication of WO2017193933A1 publication Critical patent/WO2017193933A1/zh
Priority to US16/046,716 priority patent/US10403138B2/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
    • G08G1/0175—Detecting movement of traffic to be counted or controlled identifying vehicles by photographing vehicles, e.g. when violating traffic rules
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
    • G08G1/0112—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from the vehicle, e.g. floating car data [FCD]
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/161—Decentralised systems, e.g. inter-vehicle communication
    • G08G1/162—Decentralised systems, e.g. inter-vehicle communication event-triggered
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/161—Decentralised systems, e.g. inter-vehicle communication
    • G08G1/163—Decentralised systems, e.g. inter-vehicle communication involving continuous checking
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/164—Centralised systems, e.g. external to vehicles
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/166—Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
    • G11B27/00—Editing; Indexing; Addressing; Timing or synchronising; Monitoring; Measuring tape travel
    • G11B27/02—Editing, e.g. varying the order of information signals recorded on, or reproduced from, record carriers
    • G11B27/031—Electronic editing of digitised analogue information signals, e.g. audio or video signals
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
    • G11B27/00—Editing; Indexing; Addressing; Timing or synchronising; Monitoring; Measuring tape travel
    • G11B27/10—Indexing; Addressing; Timing or synchronising; Measuring tape travel
    • G11B27/19—Indexing; Addressing; Timing or synchronising; Measuring tape travel by using information detectable on the record carrier
    • G11B27/28—Indexing; Addressing; Timing or synchronising; Measuring tape travel by using information detectable on the record carrier by using information signals recorded by the same method as the main recording
    • G11B27/30—Indexing; Addressing; Timing or synchronising; Measuring tape travel by using information detectable on the record carrier by using information signals recorded by the same method as the main recording on the same track as the main recording
    • G11B27/3081—Indexing; Addressing; Timing or synchronising; Measuring tape travel by using information detectable on the record carrier by using information signals recorded by the same method as the main recording on the same track as the main recording used signal is a video-frame or a video-field (P.I.P)
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
    • G11B27/00—Editing; Indexing; Addressing; Timing or synchronising; Monitoring; Measuring tape travel
    • G11B27/10—Indexing; Addressing; Timing or synchronising; Measuring tape travel
    • G11B27/34—Indicating arrangements 
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/698—Control of cameras or camera modules for achieving an enlarged field of view, e.g. panoramic image capture
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
    • H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources

Definitions

  • the application relates to a traffic monitoring early warning technology in the communication field, in particular to a traffic accident early warning method and a traffic accident early warning device.
  • Vehicles have become an indispensable means of transportation. In the case of a large number of vehicles and complicated road conditions, it is necessary to take measures to ensure the safety of driving.
  • various forms of terminals such as navigators, are often placed in the vehicle to remind the information of the current driving section.
  • the traffic flow status such as the degree of congestion, and whether the current road section is a high-incidence road section, to prompt the driver to pay attention to driving safety.
  • the environment around the current driving section of the vehicle, and the state of the vehicle in the traveling section, such as the number, speed, etc. of the surrounding vehicles, are dynamically changed and complicated and multi-sampled; in addition, the driver often cannot Accurate control of current traffic, especially at intersections, traffic flow, and road presence conditions such as road damage, obstacles, pedestrians, etc.
  • the embodiment of the present application provides a traffic accident early warning method and a traffic accident early warning device.
  • an embodiment of the present application provides a traffic accident warning method, including:
  • Traffic accident prediction of the target area based on a traffic scenario of the target area
  • the warning information is sent to at least one of the vehicles in the target area.
  • the embodiment of the present application provides a traffic accident early warning device, including:
  • An acquisition module configured to acquire a location of a plurality of vehicles in the interval and a video obtained by collecting the interval from at least one orientation at a location where each of the vehicles is located;
  • a parsing module configured to parse the collected video to obtain a traffic scene at a location where each of the vehicles is located
  • a synthesis module configured to synthesize a video obtained by collecting the interval from at least one orientation at a position where each of the vehicles is located, to obtain a panoramic video of the interval;
  • An identification module configured to identify a vehicle located in a target area in the panoramic video, synthesize a traffic scene of a location of a vehicle of the vehicle in the target area, to obtain a traffic scene of the target area;
  • a prediction module configured to perform the target area based on a traffic scenario of the target area Traffic accident prediction
  • the warning module is configured to send early warning information to at least one vehicle in the target area when it is determined that the traffic scene of the target area is to be a traffic accident scene.
  • an embodiment of the present application provides a traffic accident warning device, including a processor and a memory, where the processor executes a program stored in a memory to perform:
  • Traffic accident prediction of the target area based on a traffic scenario of the target area
  • the warning information is sent to at least one of the vehicles in the target area.
  • the real-time video at the location of the plurality of vehicles in the interval is obtained, the panoramic video of the interval is restored, and the real-time traffic scene of the target area is determined based on the panoramic video, thereby utilizing the traffic accident model to the traffic scene of the target area.
  • Conduct traffic accident pre-judgment so as to ensure that the relevant vehicles are alerted before traffic accidents to avoid traffic accidents. Because it is a traffic accident forecast for the actual traffic scene (ie, driving condition) of the vehicle in the interval, it is highly targeted and has high accuracy, which can significantly reduce the incidence of traffic accidents.
  • FIG. 1 is a schematic diagram of a scene of a traffic accident early warning device for traffic accident warning according to an embodiment of the present application
  • FIG. 2 is a schematic diagram of a scene of a traffic accident warning device for conducting traffic accident warning according to another embodiment of the present application
  • FIG. 3 is a schematic diagram showing the hardware structure of a traffic accident early warning device implemented as a mobile terminal according to an embodiment of the present application;
  • FIG. 4a is a schematic diagram of a vehicle-mounted mobile terminal according to an embodiment of the present application.
  • 4b is a schematic diagram of a vehicle-mounted mobile terminal according to another embodiment of the present application.
  • FIG. 5 is a schematic flow chart of a traffic accident early warning method according to an embodiment of the present application.
  • FIG. 6 is a schematic flow chart of a traffic accident early warning method according to an embodiment of the present application.
  • FIG. 7 is a schematic flowchart of an implementation manner of step 203 in FIG. 6 according to an embodiment of the present application.
  • FIG. 8 is a schematic flowchart of an implementation manner of step 204 in FIG. 6 according to an embodiment of the present application.
  • FIG. 9 is a schematic flowchart of an implementation manner of step 206 in FIG. 6 according to an embodiment of the present application.
  • FIG. 10 is a schematic flowchart of an implementation manner of step 204 in FIG. 6 according to an embodiment of the present application.
  • FIG. 11 is a schematic flow chart of a traffic accident early warning method according to an embodiment of the present application.
  • FIG. 12 is a schematic flow chart of a traffic accident early warning method according to an embodiment of the present application.
  • FIG. 13 is a diagram showing the logical function structure of a traffic accident early warning device according to an embodiment of the present application. intention
  • FIG. 14 is a schematic diagram of a traffic accident scene according to an embodiment of the present application.
  • FIG. 15 is a schematic diagram of a traffic accident warning according to an embodiment of the present application.
  • the traffic accident early warning method provided by the embodiment of the present application includes a series of steps, but the traffic accident early warning method provided by the embodiment of the present application is not limited to the described steps, and similarly, the traffic accident early warning device provided by the embodiment of the present application A series of units are included, but the traffic accident warning device provided by the embodiment of the present application is not limited to including the unit that is clearly described, and may also include a unit that is required to be set for acquiring related information or processing based on the information.
  • the embodiment of the present application provides a traffic accident early warning method to realize targeted traffic warning of a vehicle.
  • the traffic accident warning method provided by the embodiment of the present application can be applied to a traffic accident early warning device.
  • Traffic accident warning devices can be implemented in a variety of ways, the following Different implementations of the traffic accident warning device are exemplified.
  • each functional unit in the traffic accident early warning device provided by the embodiment of the present application may be distributed to the mobile terminal and the cloud implemented in the vehicle (one or more servers are deployed in the cloud as needed).
  • the mobile terminal in the interval locates the vehicle in which it is located, and performs video collection at the location of the vehicle, and transmits the vehicle position and video obtained in real time to the cloud, and the cloud is based on the real-time location of the vehicle and the location of the vehicle. Interval for traffic accident prediction and early warning.
  • an alternative scenario diagram for traffic accident warning based on the traffic accident warning device shown in FIG. 2 is shown.
  • the functional units of the traffic accident warning device may all be implemented on the mobile terminal side in the vehicle, and the mobile terminals in the interval are mutually transmitted in the real-time position of the respective set vehicles.
  • the video collected at the location of each vehicle, the mobile terminal is based on the real-time location of all the vehicles in the interval, and the video in the interval of each vehicle to warn of traffic accidents in the interval .
  • the mobile terminal can be implemented as a terminal device such as a smart phone or a tablet computer, and can be implemented in other forms, such as a notebook computer.
  • FIG. 3 exemplarily shows an optional hardware structure diagram of the foregoing mobile terminal, in particular the in-vehicle mobile terminal, including basic hardware modules such as the processor 101, the communication module 102, the memory 105, and the positioning module 103.
  • the mobile terminal can Implementing fewer or more hardware modules than Figure 3, the following examples are illustrated.
  • the hardware structure of the mobile terminal may not have a memory, a display unit, and a camera, but a data interface such as a Universal Serial Bus (USB) interface or string.
  • USB Universal Serial Bus
  • SATA Serial Advanced Technology Attachment
  • PCI Peripheral Component Interconnect
  • advanced PCI interface etc. to connect external storage devices such as flash memory, optical disk, hard disk for data storage, or
  • the data storage is performed by using the cloud via the communication module.
  • the mobile terminal can display the processing of the information or the processing result of the information by using the external display unit.
  • the display unit is only used to support the display of the information, or the built-in touch component is configured to support the touch operation.
  • the mobile terminal can use the external camera for video capture, and the mobile terminal and the external camera are connected in a wired or wireless manner to control the camera for video capture and receive the video captured by the camera.
  • a microphone can also be implemented in the mobile terminal for output of sound, and various sensors are implemented for auxiliary control of the mobile terminal.
  • the hardware structure of the mobile terminal shown in FIG. 3 is described.
  • the processor 101 is configured to execute executable instructions stored in the memory 105 (a single data processing can be performed, and the data processing result is displayed on the display unit, via the above.
  • the method for implementing the traffic accident warning method provided by the embodiment of the present application is implemented.
  • the memory 105 shown in FIG. 3 is used to store executable instructions for execution by the processor 101, intermediate results and final results of data processing by the processor 101, and data acquired from an external device to cause the processor to implement the implementation of the present application.
  • the method of warning for traffic accidents provided by the example.
  • the memory 105 may use a volatile storage medium such as a random access memory (RAM) to store intermediate processing results of data, and the memory may use a non-volatile storage medium, such as a storage device based on magnetic storage.
  • RAM random access memory
  • the memory may use a non-volatile storage medium, such as a storage device based on magnetic storage.
  • SSDs Solid State Drives
  • the communication module 102 shown in FIG. 3 is used to support communication of the mobile terminal.
  • communication The module 102 can be implemented as a cellular communication module to support the mobile terminal to access the communication network for mobile communication, and the communication system supported by the communication module can be Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA,
  • CDMA Code Division Multiple Access
  • WCDMA Wideband Code Division Multiple Access
  • the method of the present invention is not limited in the embodiment of the present application.
  • the method of the present invention is not limited.
  • the communication module can be implemented as a communication module based on various short-range communication technologies such as Wireless Complementation (WiFi), Bluetooth, and ZigBee.
  • the camera shown in FIG. 3 is used for video capture to form a video.
  • the camera may use one lens or a plurality of lenses to form a lens array to increase the viewing angle of the video capture.
  • the collected video may be in various formats such as a transport stream (TS), a program stream (PS, Program Stream), in order to enable the communication module to transmit via the communication network.
  • TS transport stream
  • PS Program Stream
  • the positioning module shown in FIG. 3 may be implemented as a positioning module based on a Global Positioning System (GPS), and output location information, speed information, and the like of the mobile terminal by receiving a GPS signal to locate the mobile terminal; alternatively, the positioning module may Implemented as a module based on other satellite positioning systems, such as the positioning module based on the China Beidou satellite positioning system, the positioning module based on the Russian GLONASS global positioning system, and the Galileo global positioning system based in Europe. Positioning module.
  • GPS Global Positioning System
  • the traffic accident warning device is implemented as a vehicle-mounted mobile terminal that is fixedly disposed inside the vehicle.
  • the in-vehicle mobile terminal can be fixedly disposed in the vehicle front panel.
  • FIG. 4b another optional schematic diagram of the vehicle-mounted mobile terminal 100 is provided in the vehicle 200.
  • the vehicle-mounted mobile terminal can also be any other position that can be fixedly set in the vehicle. Form and setting position do not do Specifically limited.
  • the in-vehicle mobile terminal has basic in-vehicle navigation functions, for example, positioning the current position of the vehicle, combining map data according to the destination indicated by the driver (mainly maintained in the in-vehicle mobile terminal, or maintained in the cloud) Calculate a feasible navigation route, start navigation according to the navigation route selected or automatically selected by the driver, prompt the vehicle's driving state (such as route, speed, etc.), and prompt the driver when the vehicle deviates from the established navigation route.
  • the in-vehicle mobile terminal has basic in-vehicle navigation functions, for example, positioning the current position of the vehicle, combining map data according to the destination indicated by the driver (mainly maintained in the in-vehicle mobile terminal, or maintained in the cloud) Calculate a feasible navigation route, start navigation according to the navigation route selected or automatically selected by the driver, prompt the vehicle's driving state (such as route, speed, etc.), and prompt the driver when the vehicle deviates from the established navigation route.
  • the vehicle's driving state such as route, speed,
  • FIG. 5 is a schematic flowchart diagram of a traffic accident warning method according to an embodiment of the present application, including the following steps:
  • Identify a vehicle located in a target area in the panoramic video synthesize a traffic scene of a location of a vehicle in the target area, and obtain a traffic scene of the target area;
  • the early warning information is sent to at least one of the vehicles in the target area.
  • the foregoing traffic accident early warning method is performed by the cloud, and the positions of the plurality of vehicles in the interval are acquired, and the location of each of the vehicles is from at least one orientation.
  • the video obtained by collecting the interval includes: receiving the location of the vehicle and the captured video transmitted by the mobile terminal provided on each vehicle.
  • the foregoing traffic accident early warning method is performed by a mobile terminal on the vehicle, and acquiring a position of the plurality of vehicles in the interval and performing the interval from the at least one orientation at a position where each of the vehicles is located
  • the captured video includes: positioning the location of the vehicle by the mobile terminal and collecting the video.
  • the traffic accident early warning device when the traffic accident early warning device is distributed in the cloud and the mobile terminal, the traffic accident processing in one section will be described.
  • the traffic accident warning in a plurality of sections can be implemented by referring to the technical solutions described below.
  • the section described in the embodiment of the present application refers to a basic area for conducting traffic accident warning.
  • the interval may be a road, a plurality of adjacent roads, an intersection, an urban area or a city, or a section may adopt a geographical area of a specific geometric shape, such as for all geographical areas where traffic accident warning is required. Square or rectangular divisions result in a series of intervals.
  • the interval may be set to be larger or smaller than the example of the foregoing interval, and the size of the interval is set according to the granularity of traffic accident handling for different geographic ranges in the actual application, and the processing capability of the cloud or mobile terminal.
  • FIG. 6 is a schematic flowchart diagram of a traffic accident early warning method according to an embodiment of the present application, including steps 201 to 207, which are described below in combination with each step.
  • Step 201 Each mobile terminal in the interval locates the location of the vehicle in which it is located, and collects a video from the location where the vehicle is located in the interval from at least one orientation.
  • the mobile terminal provided in the vehicle positions the position of the vehicle through the positioning module.
  • the location of the vehicle may be in the manner of the original geographic coordinates output by the positioning module, such as latitude and longitude coordinates, spherical coordinates, or planar two-dimensional coordinates.
  • the location of the vehicle may be in the form of grid coordinates in addition to the original geographic coordinates, in order to facilitate identification of the location of the processing vehicle in the interval.
  • Location of the vehicle The original coordinates are converted to square coordinates in the interval, and the position of the vehicle in the interval is identified by the square coordinates.
  • the grid coordinates are obtained by dividing the interval by a geographical square of a predetermined granularity.
  • the area of the geographic square is inversely proportional to the accuracy of the square coordinates. For example, the smaller the area of the geographical square, the greater the accuracy of the vehicle position represented by the square coordinates obtained by dividing the interval.
  • the size of the geographical squares used for dividing the different sections is the same, or the size of the geographic square used for dividing the different sections is inversely proportional to the traffic volume of the section. That is, the larger the traffic volume in the interval, the more busy the traffic in the interval, so the smaller traffic area is divided by the smaller geographical square, which avoids the uniform size of the interval with smaller traffic volume.
  • the division of geographic squares leads to the problem of excessive consumption of computing resources in the cloud.
  • the mobile terminal performs video capture on the interval at the location using the camera in a fixed orientation.
  • the number of cameras used is at least one, and of course the number of cameras used may be plural to form a wide viewing angle acquisition mode.
  • the camera can be placed on top of the vehicle to maximize the viewing angle of the acquisition.
  • the mobile terminal uses the camera to perform video acquisition in a manner of transforming the orientation to form a full-view acquisition mode.
  • the dynamic target in the interval at a predetermined distance e.g, 10 meters
  • the dynamic target is used as the primary acquisition target, and the camera is used within a predetermined distance from the vehicle in the interval.
  • the dynamic targets (such as traveling vehicles and pedestrians) perform tracking acquisition, wherein the mobile terminal uses binocular camera positioning technology to determine the distance between the vehicle and the dynamic target, or the mobile terminal uses the depth camera to detect the distance between the dynamic target and the vehicle.
  • Step 202 Each mobile terminal in the interval sends the vehicle location of the vehicle in which it is located, and the captured video to the cloud.
  • the video transmission is reduced in order to reduce the transmission rate of the mobile communication link.
  • the delay ensures the real-time processing of the video in the cloud, and the mobile terminal compresses the captured video and transmits it to the cloud by using video compression technology.
  • step 203 the cloud parses the collected video to obtain a traffic scene in which each vehicle is located in the interval.
  • the traffic scene at the location where the vehicle is located refers to the traffic state centered on the vehicle and within the radius of the effective collection distance of the camera provided by the vehicle.
  • step 2031 the cloud performs target extraction in the video collected by each vehicle.
  • the cloud uses target extraction technology, such as the frame difference method and background difference method currently used, to extract the target from each video frame of the video, and classify the target, for example, classified into two types of targets: static target and dynamic target, and collection.
  • target extraction technology such as the frame difference method and background difference method currently used
  • static target include various forms of traffic signs such as roads, buildings, traffic lights, lane signs, road lines, overpasses, and poles
  • dynamic targets include: vehicles, pedestrians, and other targets that are in motion.
  • Step 2032 Perform pattern recognition on the extracted target to obtain at least one of the following: a road scene of the road, a behavior type of the pedestrian, a behavior type of the vehicle, and a status of the traffic identifier.
  • the road scene includes: 1) a road segment identifier of the road, such as the XX road segment of Shennan Road; 2) a state of the road, including normal road, wet road surface, road surface damage, and road limit, for example, Construction, congestion or collapse of Section xx of Keyuan South Road; 3)
  • the types of behavior of the pedestrians include: crossing the road and waiting for the roadside; 4) the types of behavior of the vehicles include: overspeed, lane change, retrograde and emergency braking;
  • the state of the traffic sign taking traffic lights as an example, the status includes red, green and yellow lights.
  • Step 204 The cloud synthesizes a video obtained by collecting, by the vehicle in the interval, the interval from the at least one orientation, to obtain a panoramic video of the interval.
  • the panoramic video is synthesized by splicing the video collected by each vehicle at the same time in the interval, and forms a video that can be viewed from the omnidirectional section.
  • the following steps are described.
  • Step 2041 Mark the target features of each video frame in the video collected by each vehicle in the interval.
  • the target feature is obtained by performing feature extraction on the target extracted from the video in the foregoing step 2031.
  • the feature may be a color feature, a contour feature, a texture feature, any feature that may distinguish the target from other objects in the video frame, such as surrounding, a description of the target feature, or a sample of the target for different features.
  • the method is indexed, wherein the description of the target feature can be indexed by means of an identifier (such as a sequence number) or a feature vector (a component of one dimension of the feature vector is used to characterize a type of feature of the target).
  • Step 2042 Searching the video in a manner of describing the target feature or in the manner of the target feature displayed by the sample image to obtain a video frame having the same target feature.
  • each target feature of the marked index is sequentially searched in each video frame of the video to determine a video frame having the same target feature in the video.
  • Video frames with the same target feature are caused by the same target acquisition in the interval, and by determining all video frames having the same target feature, the potential connection relationship between the video frames can be determined.
  • Video 1 includes video frame 1, video frame 2, and video frame 3.
  • Video 2 includes video frame 4, video frame 5, and video frame 6.
  • target feature 1 is extracted from video frame 1
  • video frame 2 is extracted.
  • Target feature 1 and target feature 2 are extracted from video frame 3
  • target feature 2 is extracted from video frame 4
  • target feature 2 and target feature 3 are extracted from video frame 5,
  • the target feature 3 is extracted from the video frame 6; it can be seen that the video frame 1 and the video frame 2 have the same target feature 1, the video frame 2 and the video frame 3 have the same target feature 1, and the video frame 3 and the video frame 4 have the same Target feature 2, video frame 4 and video frame 5 have the same target feature 2, video frame 5 and video frame 6 have the same target feature 3, then video frame 1 to video frame 6 is a continuous acquisition of a region in the interval It is found that the areas corresponding to the video 1 to the video 6 in the interval are sequentially connected, and the potential connection relationship between the video 1 and the video 6 is as shown in Table 1.
  • the same target may be different in different mobile captured video, and therefore, optionally, in each video frame of the video
  • the target feature and the geometric deformation (such as rotation, stretching) of the target feature are matched with the target features of each video frame of the video, and the matching is matched to determine the successfully matched video.
  • Frame has phase Same target characteristics.
  • Step 2043 synthesizing the searched video frames having the same target feature.
  • the video frames are spliced based on the same target feature until all video frames are spliced to form a panoramic video frame of the interval.
  • Step 205 Identify a vehicle located in the target area in the panoramic video, perform a synthesis based on a traffic scene of an area in which each vehicle is located in the target area, and obtain a traffic scene of the target area.
  • the target region is the entire region of the entire interval
  • the corresponding region is the target region accordingly
  • the need for early warning of traffic accidents such as high-incidence road sections, road sections with high traffic flow, or intersections.
  • the warning of the target area requires obtaining a complete traffic scene of the target area.
  • the traffic scene based on the location of each vehicle in the interval has been determined. Since the video captured by a single vehicle is limited by the acquisition orientation, it cannot be based on a single The vehicle collects video to determine the complete traffic scene where the vehicle is located. In response to this situation, in one implementation, the vehicle included in the target area is identified from the panoramic video, and the traffic scene of the location of the vehicle in the target area is adjusted. Combine to form a complete traffic scene of the target area.
  • the vehicle 1 Taking the target section as an intersection as an example, the vehicle 1, the vehicle 2 and the vehicle 3 are included in the target section, wherein the traffic scene corresponding to the video collected by the vehicle 1 to the vehicle 3 is as shown in Table 2:
  • the traffic scene corresponding to the video captured by the vehicle 1 to the vehicle 3 is a one-sided traffic scene, and cannot accurately describe the traffic scene of the target area, by passing the vehicle 1 to the vehicle 3
  • the traffic scene integrated by the traffic scene corresponding to the video captured by any vehicle serves as the traffic scene of the target area, and can comprehensively describe the traffic condition of the target area.
  • Step 206 The cloud performs traffic accident prediction, and determines whether the traffic scenario of the target area will become a traffic accident scene.
  • an alternative flow diagram for determining whether the traffic scenario of the target area will become a traffic accident scenario includes steps 2061 to 2063 , and each step is described below.
  • Step 2061 Extracting a feature of a traffic accident scenario corresponding to each type of traffic accident in the traffic accident model library, where the traffic accident model library includes a correspondence between a traffic accident scenario and a traffic accident type.
  • the traffic accident model library maintained by the cloud includes a traffic accident model of different types of traffic accidents, and the traffic accident model includes a traffic accident scene and a corresponding type of traffic event. Therefore, the traffic accident scenes of different traffic accidents include three characteristics of road scene, pedestrian behavior and vehicle behavior. Of course, other types of features may also be included, such as an optional example of the traffic accident model library. Shown as follows:
  • the traffic accident model library shown in Table 3 is merely exemplary.
  • the traffic accident model has a one-to-one correspondence with the traffic accident scene.
  • the same type of traffic accident may correspond to multiple traffic accident scenarios.
  • Step 2062 Matching features of the traffic accident scene corresponding to each type of traffic accident model with characteristics of the traffic scene of the target area, and determining an accident probability of the traffic scene to evolve into a corresponding traffic accident scene based on the matching degree of the feature.
  • the degree of matching between the characteristics of the traffic accident scene corresponding to each type of traffic accident model and the characteristics of the traffic scene of the target area determines the probability of the accident that the traffic scene evolves into a corresponding traffic accident scene, Based on the quantitative relationship between the degree of matching and the probability of the accident, for example, the quantitative relationship may be a proportional relationship, or any curve having a single above trend may be used.
  • Step 2063 determining that the traffic scene of the target area will evolve into a traffic accident scene with the highest probability of accidents.
  • the highest accident probability is lower than the accident probability threshold for the following cases, which indicates that the traffic scene of the target area in the current time is likely to evolve into a traffic accident scene, so that the traffic scene of the target area is determined It is transformed into a traffic accident scene with the highest accident probability and the accident probability is higher than the accident probability threshold to ensure the accuracy of traffic accident warning.
  • the accident probability threshold can be adjusted according to the accuracy of the early warning of the vehicle driver feedback. If the accuracy of the warning reported by the driver of the vehicle is too low, the value of the accident probability threshold is small, which affects the accuracy of the warning. High (according to the established range) the probability threshold of the accident until the accuracy of the driver's feedback is up to the practical standard.
  • step 207 when the cloud determines that it will become a traffic accident scene, it sends an early warning message to the vehicle in the target area.
  • the cause of the traffic accident of the involved vehicle is analyzed according to the type of the traffic accident scene, and the vehicle is involved.
  • Sending driving prompt information, instructing the involved vehicle to follow the prompt for example, do not overspeed, do not line up to prevent traffic accidents, to avoid the traffic accident scene.
  • the traffic scene is determined by the video collected by the vehicle at the location in the interval, and the video captured by the vehicle in the interval is synthesized to restore the panoramic video of the interval, thereby
  • the video determines the vehicle in the target interval, and integrates the traffic scene based on the location of the vehicle in the target interval to obtain a complete traffic scene of the target interval, so that the traffic accident model can be used to predict the traffic accident of the complete traffic scene in the target interval. It is judged that the vehicle is warned before the traffic accident to avoid the occurrence of traffic accidents.
  • the traffic accident prediction for the actual traffic scene of the vehicle in the interval is highly targeted and has high prediction accuracy, which can significantly reduce the incidence of traffic accidents.
  • the panoramic video of the restoration interval In the panoramic video of the restoration interval, if the traffic volume in the interval is small, the panoramic video of the restoration interval completely depends on the video captured by the vehicle in the interval, and there may be a single acquisition direction and less video collected, and the coverage cannot be covered. In the entire area, it is impossible to restore the panoramic video of all areas in the section.
  • a traffic scene feature library is further set, and the video in the traffic scene feature library is used as a supplement to the use of the mobile terminal to collect video synthesized panoramic video.
  • the traffic scene feature library may include at least one of the following:
  • static targets are extracted from the video collected by each vehicle in the interval, that is, targets with static invariants, such as roads, overpasses, traffic lights, lane lines, road signs, and poles. .
  • synthesizing the static traffic scene of the interval based on the static target, and using the static traffic scene to iteratively update to a video of a static traffic scene corresponding to the interval in the traffic scene feature database, and the static traffic scene of the interval may be over time Continuous improvement until coverage of all areas of the interval.
  • a third party obtained from a third party such as the traffic management department's surveillance video database, collects surveillance video for specific surveillance areas in the interval, such as intersections, overpasses, etc., static targets extracted from surveillance video, and uses static targets.
  • the static traffic scene of the interval is synthesized, and the synthesized static traffic scene is iteratively updated into the traffic scene feature database.
  • an optional flow diagram of the panoramic video of the cloud obtaining interval in step 204 shown in FIG. 6 includes steps 2044 and 2045, and each step is described below.
  • Step 2044 obtaining a static traffic scene from the traffic scene feature library.
  • the traffic volume in the interval is lower than the traffic flow threshold, that is, the video captured at the current time in the interval cannot be completely restored to the panoramic video of the interval, and the static traffic scenario is obtained from the traffic scene feature database.
  • Step 2045 synthesizing the static traffic scene of the section obtained from the traffic scene feature library, together with the video obtained by collecting each section of the vehicle from the at least one orientation in the section, and correspondingly obtaining the section.
  • Panoramic video synthesizing the static traffic scene of the section obtained from the traffic scene feature library, together with the video obtained by collecting each section of the vehicle from the at least one orientation in the section, and correspondingly obtaining the section.
  • the target features of each video frame in the video captured by each vehicle in the interval, and the target features of each video frame in the video acquired from the traffic scene feature library are indexed to target features.
  • the video including the video captured by the camera at the location of the mobile terminal in the interval, and the at least one type of video acquired from the traffic scene feature library, in a manner described, or in a manner that the target features are displayed in the sample map Searching for video frames with the same target features, synthesizing the searched video frames with the same target features.
  • the video collected by the mobile terminal in the interval mainly constructs a dynamic traffic scene in the panoramic video
  • the static traffic scene obtained from the traffic scene feature library is mainly used to restore a still scene in the panoramic video, such as a building, etc.
  • the supplement can completely restore the panoramic video of the interval, thereby accurately determining the complete traffic scene of the target area, and ensuring the accuracy of the traffic accident prediction by using the traffic accident model to the traffic scene of the target area.
  • An optional flow diagram of the traffic accident warning method further includes the following steps:
  • Step 208 The cloud generates a new traffic accident model based on the traffic scenario of the target area and the correspondence relationship with the traffic accident type corresponding to the traffic accident scene with the highest accident probability. type.
  • Step 209 The cloud iteratively updates the traffic accident model library by using the new traffic accident model.
  • a new traffic accident model is formed by using the new traffic accident scenario and the corresponding traffic accident type, and updated to the traffic accident model feature database in an accumulated manner, in the traffic accident model library.
  • the traffic accident model follows the continuously updated traffic accident scene in the interval to realize adaptive learning. It is not necessary to manually update the traffic accident model library manually. With the continuous prediction of traffic accidents in the interval, the traffic accident model library is used for traffic. Accident prediction accuracy will be higher and higher, ensuring the accuracy of traffic accident model prediction.
  • the traffic accident warning device is a mobile terminal
  • the traffic accident warning in one section will be described.
  • the traffic accident warning in a plurality of sections can be implemented by referring to the technical solutions described below.
  • the traffic accident warning processing shown in FIG. 12 is only performed cooperatively by the mobile terminal, and does not require participation of the cloud.
  • FIG. 12 is a schematic flowchart diagram of a traffic accident early warning method according to an embodiment of the present application, including steps 301 to 307, which are described below in conjunction with the steps.
  • Step 301 Each mobile terminal on the vehicle in the interval locates the location of the vehicle in which it is located, and collects a video from the location where the vehicle is located in the interval from at least one orientation.
  • the location of the vehicle may take the form of original geographic coordinates or geographic square coordinates.
  • information of different intervals may be preset in the mobile terminal, and the mobile terminals open the positioning service between each other so that any mobile terminal can acquire the real-time location of other mobile terminals in the interval.
  • Step 302 Each mobile terminal in the interval sends the location of the vehicle in which it is located, and the captured video to other mobile terminals.
  • the mobile terminal can compress the collected video in consideration of different data processing capabilities of different mobile terminals. Send to ensure that all mobile terminals have enough space to receive video and process it.
  • Step 303 The mobile terminal in the interval parses the collected video to obtain a traffic scene in which the vehicle is located in the interval.
  • each vehicle in the interval parses the captured video and the video transmitted by other mobile terminals in the interval to obtain a traffic scenario in which each mobile terminal is located in the interval.
  • the traffic scene in which the mobile terminal analyzes the video to obtain the location in the interval may be implemented by referring to the traffic scene collected by the cloud analysis to obtain the traffic scene of the location of each vehicle in the interval.
  • Example 2 In another implementation manner, the vehicle in the interval performs capability negotiation to determine at least one mobile terminal with the highest processing capability in the interval as a node of the interval, and parses the video collected by all the vehicles in the interval to obtain the vehicle in the interval. Location of the traffic scene.
  • Step 304 Each mobile terminal in the interval synthesizes a video obtained by collecting, for each vehicle in the interval, from the at least one orientation, and correspondingly obtaining a panoramic video of the interval.
  • Each mobile terminal in the section performs panoramic synthesis based on the acquired video and the received video, and the synthesized panoramic video can be implemented by referring to the description of the aforementioned cloud-composited panoramic video.
  • the mobile terminal as a node performs panoramic panoramic video synthesis based on the acquired video and the received video.
  • Step 305 The mobile terminal in the interval identifies the vehicle located in the target area in the panoramic video, and performs synthesis based on the traffic scene of the area in which each vehicle is located in the target area, to obtain a traffic scene of the target area.
  • each mobile terminal in the interval identifies a vehicle located in a target area in the panoramic video, based on a traffic scene of an area in which each vehicle is located in the target area Synthesize to obtain a traffic scene of the target area.
  • the traffic scene identifying the target area can refer to the traffic of the aforementioned cloud recognition target area.
  • the mobile terminal as a node in the section identifies the vehicle located in the target area in the panoramic video, based on the traffic of the area in which each vehicle is located in the target area.
  • the scene is synthesized to obtain a traffic scene of the target area.
  • Step 306 The mobile terminal in the interval determines whether the traffic scene of the target area will become a traffic accident scene.
  • each mobile terminal in the interval determines whether the traffic scene of the target area will become a traffic accident scene. Similarly, it is determined whether or not the traffic scene of the target area will become a traffic accident scene by referring to the cloud to determine whether the traffic scene of the target area will become a traffic accident scene.
  • the mobile terminal as a node in the section determines whether the traffic scene of the target area will become a traffic accident scene.
  • Step 307 The mobile terminal sends an early warning message to the vehicle in the target area when it determines that it will become a traffic accident scene.
  • the mobile terminal when it is determined that the traffic scene of the area in which it is located will become a traffic accident scene, the mobile terminal issues an early warning and issues an early warning to the mobile terminal in other vehicles involved in the traffic accident.
  • a schematic diagram of a traffic accident early warning device 400 includes: an obtaining module 401, a parsing module 402, a synthesizing module 403, and identification. Module 404, prediction module 405 and early warning module 406.
  • the traffic accident warning device 400 can also include a traffic accident model library 407 and a traffic scene feature library 408. Each module will be described below.
  • the obtaining module 401 is configured to acquire a position of a plurality of vehicles in the interval and a video obtained by collecting the interval from the at least one orientation at a position where each of the vehicles is located.
  • the acquisition module 401 when the traffic accident warning device 400 is implemented in a mobile terminal, the acquisition module 401 includes a positioning unit and an acquisition unit.
  • the positioning unit may be implemented by using the positioning module 103 shown in FIG. 3, and the positioning module 103 receives the GPS signal or the Beidou positioning signal to locate the vehicle.
  • the position of the vehicle may be in the geographical position instead of the original geographical coordinates.
  • the acquisition unit is configured to collect the video from the at least one orientation at a location where the vehicle is located to obtain a video.
  • the acquisition unit can be implemented by the camera 106 shown in FIG. 3, and the acquisition unit can be implemented as a camera or as a module composed of multiple cameras 106 for full-scale video capture.
  • the acquisition unit may also integrate the pan-tilt device to adjust the acquisition orientation of the camera 106.
  • the acquisition unit is placed on top of the vehicle so that the angle of view of the acquisition can be maximized.
  • the parsing module 402 is configured to parse the collected video to obtain a traffic scene in which each vehicle in the interval is located.
  • the parsing module 402 performs target extraction in the video collected by each vehicle, and obtains at least one of the following objectives: road, vehicle, pedestrian, and traffic light;
  • the target performs pattern recognition to obtain at least one of the following: a road scene of the road, a behavior type of the pedestrian, a behavior type of the vehicle, and a status of the traffic light.
  • the synthesizing module 403 synthesizes a video obtained by collecting each of the vehicles in the interval from at least one orientation, and correspondingly obtaining a panoramic video of the interval.
  • the identification module 404 identifies the vehicle located in the target area of the panoramic video, and performs a synthesis based on the traffic scene of the area in which the vehicle is located in the target area to obtain a traffic scene of the target area.
  • the synthesizing module 403 tags the target features of each video frame in the video captured by each vehicle in the interval; Displayed as a description of the target feature, or as a sample The target feature is searched for the video to obtain video frames having the same target feature; the searched video frames having the same target feature are synthesized.
  • the prediction module 405 determines whether the traffic scene of the target area will become a traffic accident scene.
  • the prediction module 405 extracts features of a traffic accident scenario corresponding to each type of traffic accident model in the traffic accident model library 407, the traffic accident
  • the model library includes a correspondence relationship between the traffic accident scene and the type of traffic accident; matching the characteristics of the traffic accident scene corresponding to each type of traffic accident with the characteristics of the traffic scene of the target area, and determining the evolution of the traffic scene based on the degree of matching of the features The probability of accidents for the corresponding traffic accident scene; determining the traffic scene of the target area will evolve into the traffic accident scene with the highest probability of accidents.
  • the early warning module 406 issues warning information to the vehicles in the target area when the prediction module 405 determines that it will become a traffic accident scene.
  • the early warning module 406 performs at least one of: alerting a vehicle involved in the traffic accident scene in the target area with a traffic accident of a type corresponding to a traffic accident scene;
  • the vehicle involved in the traffic accident scene issues driving prompt information for instructing the involved vehicle to follow the prompt to avoid the traffic accident scene.
  • the parsing module 402, the synthesizing module 403, the identifying module 404, the predicting module 405, and the warning module 406 are divisions of the traffic accident warning device at the logical function level, and any of the modules may be implemented in combination, or any of them.
  • the module may be implemented in a manner of splitting a plurality of modules, and the parsing module 402, the synthesizing module 403, the identifying module 404, the predicting module 405, and the alerting module 406 may each be executed by the processor 101 shown in FIG. 3 by executing executable instructions stored in the memory 105. achieve.
  • the parsing module 402 extracts a static target in the video collected by each vehicle, synthesizes the static traffic scene of the interval based on the static target, and iteratively updates to the corresponding corresponding in the traffic scene feature library 406 by using the static traffic scene.
  • a static traffic scenario of the interval or, acquiring, from a third-party monitoring video database, a monitoring video collected by a third party for a specific monitoring area in the interval, and updating the monitoring video to the corresponding interval in the traffic scene feature library 406 Monitoring video.
  • the synthesizing module is further configured to obtain at least one of the following: a monitoring video of a third party for a specific monitoring area in the interval; a video corresponding to the static traffic scene in the interval; and the traffic scene from the traffic scene
  • At least one type of video acquired by the feature library together with a video obtained by collecting, for each of the vehicles in the interval from the at least one orientation, correspondingly obtains a panoramic video of the interval.
  • the video of the static scene in the static scene feature library is used to make up for the difference of the static scenes in the video captured by the vehicle, and the synthesized panoramic video includes the complete traffic scene.
  • the early warning module 406 in order to allow the traffic accident model in the traffic accident model library to be adaptively updated during the prediction process of the traffic accident to enable comprehensive and accurate determination of the diverse traffic accident scenarios, the early warning module 406, And a method for generating a new traffic accident model based on a traffic scenario of the target area and a correspondence relationship of a traffic accident type corresponding to the traffic accident scene with the highest accident probability; and iteratively updating the new traffic accident model Traffic Accident Model Library 407.
  • an optional schematic diagram of the target area in the section shown in FIG. 14 is taken as an example of the target area in the section as an intersection.
  • the vehicle B travels from north to south in the north section of the intersection
  • the vehicle A travels from east to west at the east section of the intersection
  • the vehicle A and the vehicle A are limited by the driver's perspective of the vehicle A and the vehicle B.
  • the driver of the vehicle B cannot perceive the presence of the other party, and once the vehicle A is running with a red light, an accident of collision with the vehicle B is inevitable.
  • Video capture and positioning of the location of the vehicle, the cloud according to the video uploaded by the mobile terminal of the vehicle A, the vehicle B, and the panoramic video of the location synthesis intersection can also combine the traffic scene feature library with the video of the intersection of the traffic scene feature library when synthesizing the panoramic video The compositing is performed to restore the panoramic video of the complete traffic scene at the intersection of the current moment.
  • the cloud resolves the traffic scene at the intersection to the panoramic video, as shown in Table 4:
  • the traffic accident model in the traffic accident model library is used and the traffic accident prediction based on the panoramic video is matched to the traffic accident model where the two vehicles collide at the intersection, as shown in Table 5:
  • the mobile terminal can display a prompt message on the display screen and issue a buzzer alarm prompt.
  • the driver of the vehicle A prompts that the red light is about to crash
  • the driver of the vehicle B prompts that the vehicle is in the other direction. Take a red light and pay attention to avoiding.
  • An embodiment of the present application further provides a computer storage medium, wherein the storage is executable A line instruction for executing the aforementioned traffic accident warning method.
  • the real-time video of the location of the vehicle in the interval is collected, and the panoramic video of the interval is restored, and the real-time traffic scene of the target area can be accurately determined based on the panoramic video, thereby using the traffic accident model to target
  • the traffic scene of the interval is pre-judged for traffic accidents, so that the relevant vehicles are pre-warned before the traffic accident to avoid the occurrence of traffic accidents, because the traffic accidents are predicted for the actual traffic scenes (ie, driving conditions) of the vehicles in the interval. Targeted and accurate, it can significantly reduce the incidence of traffic accidents.
  • the integrated modules described in the embodiments of the present application may also be stored in a computer readable storage medium if they are implemented in the form of software functional modules and sold or used as separate products. Based on such understanding, those skilled in the art will appreciate that embodiments of the present application can be provided as a method, system, or computer program product. Thus, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment in combination of software and hardware.
  • the application can take the form of a computer program product embodied on one or more computer-usable storage media containing computer usable program code, including but not limited to a USB flash drive, a mobile hard drive, a read only memory (ROM, Read-Only Memory), Random Access Memory (RAM), disk storage, CD-ROM, optical storage, and the like.
  • a USB flash drive a mobile hard drive
  • ROM read only memory
  • RAM Random Access Memory
  • disk storage CD-ROM, optical storage, and the like.
  • These computer program instructions can also be stored at a computer or other programmable data location.
  • the device is readable in a computer readable memory that operates in a particular manner such that instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device implemented in a flow or a flow and/or block diagram of the flowchart The function specified in the box or in multiple boxes.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Signal Processing (AREA)
  • Theoretical Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Atmospheric Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Chemical & Material Sciences (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种交通事故预警方法及交通事故预警装置；方法包括：获取区间中多个车辆的位置以及在每个车辆所处位置处从至少一个方位对所述区间进行采集得到的视频(101)；解析所采集到的视频，得到每个车辆所处位置的交通场景(102)；合成在每个车辆所处位置处从至少一个方位对区间进行采集得到的视频，得到区间的全景视频(103)；识别出位于全景视频中的目标区域中的车辆，对目标区域中的车辆的车辆所处位置的交通场景进行合成，得到目标区域的交通场景(104)；基于目标区域的交通场景对目标区域进行交通事故预测(105)；以及当判断目标区域的交通场景将成为交通事故场景时，向目标区域中的至少一个车辆发出预警信息(106)。

Description

交通事故预警方法及交通事故预警装置
本申请要求于2016年05月10日提交中国专利局、申请号为201610308302.4、发明名称为“交通事故处理方法及交通事故处理装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信领域的交通监控预警技术，尤其涉及一种交通事故预警方法及交通事故预警装置。
背景技术
车辆已经成为不可缺少的交通工具。在车辆多、路况复杂的情况下，有必要采取措施保证行车安全。
目前，往往是在车辆中设置各种形式的终端，如导航仪等，对当前行驶路段的信息进行提醒。例如提醒车辆当前所行驶路段的限速信息、车流状况，如拥堵程度，以及当前路段是否是事故高发路段，以提示驾驶者注意行车安全。
但是，车辆当前行驶路段周围的环境、以及所行驶路段中的车辆的状态，如，周围的车辆的数量、速度等，是动态变换且复杂多采样的；另外，因为存在视野死角驾驶者往往不能对当前的行车进行准确控制，尤其是在交叉路口、车流量大以及路面存在状况，如路面破损、障碍物、行人等的情况下。
发明内容
本申请实施例提供了一种交通事故预警方法及交通事故预警装置。
第一方面，本申请实施例提供一种交通事故预警方法，包括：
获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频；
解析所采集到的视频，得到每个所述车辆所处位置的交通场景；
合成在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频；
识别出位于所述全景视频中的目标区域中的车辆，对所述目标区域中的车辆的车辆所处位置的交通场景进行合成，得到所述目标区域的交通场景；
基于所述目标区域的交通场景对所述目标区域进行交通事故预测；以及
当判断所述目标区域的交通场景将成为交通事故场景时，向所述目标区域中的至少一个车辆发出预警信息。
第二方面，本申请实施例提供一种交通事故预警装置，包括：
获取模块，用于获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频；
解析模块，用于解析所采集到的视频，得到每个所述车辆所处位置的交通场景；
合成模块，用于合成在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频；
识别模块，用于识别出位于所述全景视频中的目标区域中的车辆，对所述目标区域中的车辆的车辆所处位置的交通场景进行合成，得到所述目标区域的交通场景；
预测模块，用于基于所述目标区域的交通场景对所述目标区域进行 交通事故预测；以及
预警模块，用于当判断所述目标区域的交通场景将成为交通事故场景时，向所述目标区域中的至少一个车辆发出预警信息。
第三方面，本申请实施例提供一种交通事故预警装置，包括处理器和存储器，所述处理器执行存储器中存储的程序以执行：
获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频；
解析所采集到的视频，得到每个所述车辆所处位置的交通场景；
合成在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频；
识别出位于所述全景视频中的目标区域中的车辆，对所述目标区域中的车辆的车辆所处位置的交通场景进行合成，得到所述目标区域的交通场景；
基于所述目标区域的交通场景对所述目标区域进行交通事故预测；以及
当判断所述目标区域的交通场景将成为交通事故场景时，向所述目标区域中的至少一个车辆发出预警信息。
本申请实施例中，通过获取区间中多个车辆所在位置处的实时视频，还原出区间的全景视频，并基于全景视频确定目标区域的实时交通场景，从而利用交通事故模型对目标区域的交通场景进行交通事故预判，这样就可以保证在交通事故发生前对相关车辆进行预警，以避免交通事故的发生。由于是针对区间中车辆的实际交通场景(也就是行车状况)进行交通事故的预测，针对性强且准确率高，能够显著降低交通事故的发生率。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据提供的附图获得其他的附图。
图1为根据本申请一实施例的交通事故预警装置进行交通事故预警的一个场景示意图；
图2为根据本申请另一实施例的交通事故预警装置进行交通事故预警的一个场景示意图；
图3为根据本申请一实施例的实施为移动终端的交通事故预警装置的硬件结构示意图；
图4a为根据本申请一实施例的车载移动终端的示意图；
图4b为根据本申请另一实施例的车载移动终端的示意图；
图5为根据本申请一实施例的交通事故预警方法的流程示意图；
图6为根据本申请一实施例的交通事故预警方法的流程示意图；
图7为根据本申请一实施例的图6中步骤203的实现方式的流程示意图；
图8为根据本申请一实施例的图6中步骤204的实现方式的流程示意图；
图9为根据本申请一实施例的图6中步骤206的实现方式的流程示意图；
图10为根据本申请一实施例的图6中步骤204的实现方式的流程示意图；
图11为根据本申请一实施例的交通事故预警方法的流程示意图；
图12为根据本申请一实施例的交通事故预警方法的流程示意图；
图13为根据本申请一实施例的交通事故预警装置的逻辑功能结构示 意图；
图14为根据本申请一实施例的交通事故场景的示意图；
图15为根据本申请一实施例的交通事故预警的示意图。
具体实施方式
以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所提供的实施例仅仅用以解释本申请，并不用于限定本申请。另外，以下所提供的实施例是用于实施本申请的部分实施例，而非提供实施本申请的全部实施例，在本领域技术人员不付出创造性劳动的前提下，对以下实施例的技术方案进行重组所得的实施例、以及基于对申请所实施的其他实施例均属于本申请的保护范围。
需要说明的是，在本申请实施例中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的方法或者装置不仅包括所明确记载的要素，而且还包括没有明确列出的其他要素，或者是还包括为实施方法或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的方法或者装置中还存在另外的相关要素(例如方法中的步骤或者装置中的单元)。例如，本申请实施例提供的交通事故预警方法包含了一系列的步骤，但是本申请实施例提供的交通事故预警方法不限于所记载的步骤，同样地，本申请实施例提供的交通事故预警装置包括了一系列单元，但是本申请实施例提供的交通事故预警装置不限于包括所明确记载单元，还可以包括为获取相关信息、或基于信息进行处理时所需要设置的单元。
本申请实施例提供一种交通事故预警方法以实现对车辆有针对性地进行交交通事故预警。本申请实施例提供的交通事故预警方法可以应用于交通事故预警装置。交通事故预警装置可以采用多种方式来实施，以下对 交通事故预警装置的不同实现方式进行示例性说明。
交通事故预警装置的实现方式1)
在一个实施例中，参见图1示出的基于交通事故预警装置进行交通事故预警的一个可选的场景示意图。为了节省移动终端的计算资源和存储资源，本申请实施例提供的交通事故预警装置中各功能单元可以分布实施在车辆中的移动终端和云端(根据需要，云端部署有一个或多个服务器)，其中区间中的移动终端对其所处的车辆进行定位、以及在车辆所处位置处进行视频采集，将实时获取的车辆位置以及视频发送至云端，云端基于车辆的实时位置和车辆所处位置对区间进行交通事故预测以及预警。
交通事故预警装置的实现方式2)
在另一个实施例中，参见图2示出的基于交通事故预警装置进行交通事故预警的一个可选的场景示意图。在移动终端的计算资源和存储资源足够使用的情况下，交通事故预警装置的功能单元可以全部实施在车辆中的移动终端侧，区间中的移动终端之间互相发送在各自所设置车辆的实时位置、以及在各车辆所处位置采集的视频，移动终端(每个移动终端或者部分移动终端)基于区间中的全部车辆的实时位置、以及各车辆所处位置的视频对区间中的交通事故进行预警。其中，移动终端可以实施为智能手机和平板电脑等终端设备，当然也可以实施为其他形式的移动终端，例如笔记本电脑，本申请实施例中对此不作限定。
图3示例性示出了前述移动终端特别是车载移动终端的一个可选的硬件结构示意图，包括处理器101、通信模块102、存储器105和定位模块103等基本的硬件模块，当然，移动终端可以实施相较与图3更少的或更多的硬件模块，以下举例说明。
例如，移动终端的硬件结构中可以不具有存储器、显示单元和摄像头，而是利用数据接口如通用串行总线(USB，Universal Serial Bus)接口、串 行高级技术附件(SATA，Serial Advanced Technology Attachment)接口、外设部件互连标准(PCI，Peripheral Component Interconnect)接口、高级PCI接口等连接外部的存储装置如闪存、光盘、硬盘以进行数据存储，或者经由通信模块利用云端进行数据存储。
同样地，移动终端可以利用外部的显示单元显示信息的处理过程或者信息的处理结果，可选地，显示单元仅用于支持信息的显示，或者内置触控组件以支持触控操作，本申请实施例中对此不做限定；同样地，移动终端可以利用外部的摄像头进行视频采集，移动终端与外部的摄像头以有线或无线的方式连接以控制摄像头进行视频采集并接收摄像头采集的视频。
再例如，根据实际需要，移动终端中还可以实施麦克风以用于声音的输出，实施各种传感器以用于对移动终端的辅助控制。
接续对图3示出的移动终端的硬件结构进行说明，处理器101用于执行存储器105中存储的可执行指令(可以采用单一的以进行数据处理，将数据处理结果在显示单元显示，经由上述方式来来实施本申请实施例提供的交通事故预警方法。
图3中示出的存储器105用于存储供处理器101执行的可执行指令、处理器101进行数据处理的中间结果和最终结果、以及从外部设备获取的数据，以使处理器实施本申请实施例提供的交通事故预警方法。示例性地，存储器105可以采用易失性存储介质如随机存取存储器(RAM，Random Access Memory)来存储数据的中间处理结果，存储器可以采用非易失性存储介质，如基于磁性存储的存储装置如机械硬盘、磁带，基于闪存的存储装置如固态硬盘(SSD，Solid State Drives)或其他任意形式的基于闪存的存储装置。
图3中示出的通信模块102用于支持移动终端的通信。例如，通信 模块102可以实施为蜂窝通信模块以支持移动终端接入通信网络进行移动通信，通信模块支持的通信制式可为码分多址(CDMA，Code Division Multiple Access))、宽带码分多址(WCDMA，Wideband Code Division Multiple Access)、时分-同步码分多址(TD-SCDMA，Time Division-Synchronous Code Division Multiple Access)及其演进制式，本申请实施例中对比不做限定。再例如，通信模块可以实施为基于无线相容性认证(WiFi，Wireless Fidelity)、蓝牙(Bluetooth)和紫蜂(ZigBee)等各种近距离通信技术的通信模块。
图3中示出的摄像头用于进行视频采集形成视频，示例性地，摄像头可以采用一颗镜头，或者采用多颗镜头形成镜头阵列的方式以增大视频采集的视角。可选地，采集得到的视频可以为各种格式如传送流(TS，Transport Stream)、节目流(PS，Program Stream)，以便于使通信模块经由通信网络进行传输。
图3中示出的定位模块可以实施为基于全球卫星定位系统(GPS)的定位模块，通过接收GPS信号来定位移动终端而输出移动终端的位置信息、速度信息等；可替换地，定位模块可以实施为基于其他的卫星定位系统的模块，如基于中国北斗卫星定位系统的定位模块、基于俄罗斯的格洛纳斯(GLONASS)全球定位系统的定位模块，以及基于欧洲的伽利略(Galileo)全球定位系统的定位模块。
特别地，在一个实施例中，交通事故预警装置实施为车辆内部固定设置的车载移动终端。示例性地，参见在图4a示出的在车辆200内部设置车载移动终端100的一个可选的示意图，车载移动终端可以固定设置在车辆前置面板中。当然，参见图4b示出的在车辆200内部设置车载移动终端100的另一个可选的示意图，车载移动终端也可以为车辆的其他任意可以固定设置的位置，本申请实施例中对车载移动终端的形态以及设置位置不做 具体限定。
接续对车载移动终端进行说明，车载移动终端具备基本的车载导航功能，例如，对车辆的当前位置进行定位，根据驾驶者指示的目的地结合地图数据(在车载移动终端本地维护，或者在云端维护)计算可行的导航路线，根据驾驶者选定或自动选定的导航路线开启导航，对车辆的行驶状态(如路线、速度等)进行提示，以及，在车辆偏离既定的导航路线时提示驾驶者纠正行驶路线，并在预满足预定条件时，如车辆偏离既定导航路线预定距离或者驾驶者确定驾驶车辆接续沿当前道路行驶时重新计算导航路线并接续导航，或者，车辆偏离既定的导航路线时自动结束导航。
图5示出了根据本申请一实施例的交通事故预警方法的流程示意图，包括以下步骤：
101，获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频；
102，解析所采集到的视频，得到每个所述车辆所处位置的交通场景；
103，合成在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频；
104，识别出位于所述全景视频中的目标区域中的车辆，对所述目标区域中的车辆的车辆所处位置的交通场景进行合成，得到所述目标区域的交通场景；
105，基于所述目标区域的交通场景对所述目标区域进行交通事故预测；以及
106，当判断所述目标区域的交通场景将成为交通事故场景时，向所述目标区域中的至少一个车辆发出预警信息。
在一种实现方式中，上述交通事故预警方法由云端执行，则获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所 述区间进行采集得到的视频包括：接收各车辆上设置的移动终端发送的车辆位置和所采集的视频。
在另一种实现方式中，上述交通事故预警方法由车辆上的移动终端执行，则获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频包括：由移动终端定位所处车辆的位置并进行视频的采集。
以下结合图6至图11，对交通事故预警装置分布实施在云端和移动终端时，对一个区间中的交通事故处理进行说明。对于多个区间中的交通事故预警可以参照以下记载的技术方案而实施。
本申请实施例中记载的区间是指进行交通事故预警的基本区域。示例性地，区间可以为一条道路、多条相邻的道路、十字路口、一个城区或一个城市，又或者区间可以采用特定几何形状的地理区域，如对需要进行交通事故预警的全部地理区域以方形或长方形划分得到一系列区间。当然，区间可以设置为较前述区间的示例更大或更小的范围，区间的大小根据实际应用中对不同地理范围进行交通事故处理的粒度、以及云端或移动终端的处理能力设置。
图6示出了根据本申请一实施例提供的交通事故预警方法的流程示意图，包括步骤201至步骤207，以下结合各步骤进行说明。
步骤201，区间中的每个移动终端定位其所处的车辆的位置，并从至少一个方位对车辆在区间中所处的位置进行采集得到视频。
在一个实施方式中，车辆中设置的移动终端通过定位模块定位车辆的位置。例如，车辆的位置可以采用定位模块输出的原始的地理坐标的方式，例如经纬度坐标、球坐标或平面二维坐标的方式。
再例如，车辆的位置除了采用原始的地理坐标的方式，为了便于识别处理车辆在区间中的位置，还可以采用方格坐标的形式。将车辆位置的 原始坐标转换为区间中的方格坐标，以方格坐标标识车辆在区间中的位置。方格坐标是采用预定粒度的地理方格对区间进行划分得到。地理方格的面积与方格坐标的精度成反比，例如地理方格的面积越小，则对区间进行划分得到的方格坐标表征的车辆位置的精度就越大。
可选地，车辆的位置采用方格坐标时，对不同区间进行划分所使用的地理方格的大小一致，或者，不同区间进行划分所使用的地理方格的大小与区间的车流量成反比，也就是区间中的车流量越大，表明区间的交通越繁忙，因此较车流量较小的区间采用更小的地理方格对区间进行划分，避免了对车流量较小的区间也采用统一大小的地理方格进行划分，导致云端的计算资源多度消耗的问题。
在一个实施方式中，移动终端利用摄像头采用固定方位的方式在所处位置对区间进行视频采集。所利用的摄像头的数量至少为一个，当然所利用的摄像头的数量可以为多个从而形成广视角的采集模式。特别地，摄像头可以设置在车辆的顶部从而最大限度增大采集的视角。
在另一个实现方式中，移动终端利用摄像头对采用变换方位的方式进行视频采集以形成全视角的采集模式。特别地，由于区间中与车辆处于预定距离(如10米)的动态目标往往是导致交通事故的潜在目标，相应地，将动态目标作为首要的采集对象，利用摄像头对区间中与车辆预定距离内的动态目标(如行驶的车辆和行人)进行跟踪采集，其中，移动终端采用双目摄像头定位技术确定车辆与动态目标的距离，或者，移动终端利用深度摄像头探测动态目标与车辆的距离。
步骤202，区间中的每个移动终端将其所处车辆的车辆位置、以及采集得到视频发送至云端。
在一个实现方式中，考虑到移动终端采用移动通信的方式向云端发送所采集的视频，为了减少移动通信链路的传输速率有限对视频传输带来 的延时，确保云端对视频进行处理的实时性，移动终端采用视频压缩技术对所采集的视频进行压缩并传送至云端。
步骤203，云端解析所采集到的视频，得到区间中每个车辆所处位置的交通场景。
车辆所处位置的交通场景是指以车辆为中心，在车辆设置的摄像头的有效采集距离为半径的范围内的交通状态。在一个实现方式中，参见图7示出的云端解析所采集到的视频得到所述区间中每个车辆所处位置的交通场景的一个流程示意图，包括步骤2031至步骤2032，以下对各步骤进行说明。
步骤2031，云端在每个车辆所采集到的视频中进行目标提取。
云端对采用目标提取技术，例如目前采用的帧差法、背景差法等，从视频的各视频帧中提取得到目标，对目标进行分类，例如分类为静态目标和动态目标两类目标，与采集到视频使车辆的位置叠加，得到车辆在相应位置所采集到的目标。示例性地，静态目标包括：道路、建筑物、交通灯、车道牌、道路线、过街天桥和立杆等各种形式的交通标识；动态目标包括：车辆、行人和其他任意处于运动状态的目标。
步骤2032，对所提取的目标进行模式识别，得到以下特征至少之一：所述道路的道路场景、所述行人的行为类型、所述车辆的行为类型和所述交通标识的状态。
示例性地，所述道路场景包括：1)所述道路的路段标识，如深南大道xx路段；2)所述道路的状态，包括道路正常、路面湿滑、路面破损和道路限行，例如，科苑南路xx路段施工、拥堵或者塌陷；3)所述行人的行为类型包括：横穿道路和路边等待；4)所述车辆的行为类型包括：超速、更换车道、逆行和紧急刹车；5)交通标识的状态，以交通灯为例，状态包括红灯、绿灯和黄灯。
步骤204，云端合成所述区间中每个车辆从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频。
全景视频是对区间中各车辆在同一时刻采集的视频进行拼接的方式合成，形成从全方位可以对区间进行观看的视频。在一个实现方式中，参见图8示出的合成全景视频的一个流程示意图，包括步骤2041至步骤2043，以下对各步骤进行说明。
步骤2041，对所述区间中每个车辆采集得到的视频中各视频帧的目标特征进行标记索引。
目标特征是前述的步骤2031中对从视频中所提取的目标进行特征抽取得到。示例性地，特征可以为颜色特征、轮廓特征、纹理特征、与周围等任意可以将目标从视频帧中的其他目标区分开的特征，对于不同的特征采用目标特征的描述、或者目标的样图的方式进行索引，其中目标特征的描述可以采用标识(如序号)或者特征向量的方式进行索引(特征向量的一个维度的分量用于表征目标的一种类型的特征)。
步骤2042，以目标特征的描述的方式、或以样图显示的目标特征的方式对所述视频进行搜索得到具有相同目标特征的视频帧。
对于区间中各移动终端在同一时刻采集的视频，在视频的每个视频帧中依次搜索所标记索引的每个目标特征，确定视频中具有相同目标特征的视频帧。具有相同目标特征的视频帧是因为对区间中同一目标采集导致，通过确定所有具有相同目标特征的视频帧，可以确定各视频帧之间的潜在的连接关系。
例如，假设视频1包括视频帧1、视频帧2和视频帧3，视频2包括视频帧4、视频帧5和视频帧6，如果视频帧1中提取出目标特征1，视频帧2中提取出目标特征1，视频帧3中提取出目标特征1和目标特征2，视频帧4中提取出目标特征2，视频帧5中提取出目标特征2和目标特征3， 视频帧6中提取出目标特征3；可见，视频帧1与视频帧2具有相同的目标特征1，视频帧2与视频帧3具有相同的目标特征1，视频帧3与视频帧4具有相同的目标特征2，视频帧4与视频帧5具有相同的目标特征2，视频帧5与视频帧6具有相同的目标特征3，则视频帧1至视频帧6是对区间中一区域进行连续的采集得到，视频1至视频6在区间中所对应的区域是依次连接的，视频1至视频6之间存在潜在的连接关系如表1所示。
Figure PCTCN2017083762-appb-000001
表1
需要指出的是，由于不同移动终端利用摄像头进行采集的方位不可能完全相同，因此同一目标在不同移动采集的视频中会有所区别，因此，可选地，在视频的每个视频帧中依次搜索所标记索引的每个目标特征时，将目标特征以及目标特征的几何变形(如旋转、拉伸)与视频的每个视频帧所具有的目标特征进行匹配，匹配一致则判定匹配成功的视频帧具有相 同的目标特征。
步骤2043，对所搜索得到的具有相同目标特征的视频帧进行合成。
在一个实现方式中，对具有相同目标特征的视频帧，以相同的目标特征为基准对视频帧进行拼接，直至拼接完毕所有的视频帧，形成区间的全景视频帧。
例如，接续前述视频帧1至视频帧6的示例进行说明，视频帧1至视频帧6之间基于目标特征的连接关系如下：视频帧1-(目标特征1)>视频帧2-(目标特征1)>视频帧3-(目标特征2)>视频帧4-(目标特征2)>视频帧5-(目标特征3)>视频帧6，将视频帧1与视频帧2基于目标特征1进行拼接合成，将视频帧2与视频帧3基于目标特征1进行拼接合成，将视频帧3与视频帧4基于目标特征2进行拼接合成，将视频帧4与视频帧5基于目标特征2进行拼接合成，将视频帧5与视频帧6基于目标特征3进行拼接合成，形成视频帧1至视频帧6拼接合成的全景视频。
步骤205，识别出位于所述全景视频中目标区域的车辆，基于所述目标区域中每个车辆所处区域的交通场景进行合成，得到所述目标区域的交通场景。
对于一个区间，存在对区间的全部区域(相应地，目标区域为整个区间的全部区域)的交通场景进行交通事故进行预警的需求，或者存在对区间中的特定区域(相应地特定区域为目标区域，如事故高发路段、车流量较高的路段、或者是十字路口等)进行交通事故进行预警的需求。对目标区域预警需要获取目标区域的完整的交通场景，在前述步骤203中已经确定基于区间中每个车辆所处位置的交通场景，由于单个车辆采集的视频由于采集方位的局限，因此无法根据单个车辆采集视频确定车辆所处位置的完整的交通场景。针对此情况，在一个实现方式中，从全景视频中识别出目标区域所包括的车辆，对目标区域中车辆所处位置的交通场景进行整 合，形成目标区域的完整的交通场景。
以目标区间为十字路口为例，在该目标区间中包括车辆1、车辆2和车辆3，其中，车辆1至车辆3采集的视频对应的交通场景如表2所示：
  道路场景 车辆行为 行人行为
车辆1 **北路、北向畅通 车辆3并线 无行人
车辆2 **南路，路面坑洼 无其他车辆 有2行人在等待
车辆3 **北路，北向畅通 车辆1靠右行驶 无行人
表2
从表2可以看出，采用车辆1至车辆3任一车辆采集的视频对应的交通场景作为目标区间的交通场景都是片面的，不能准确描述目标区域的交通场景，通过将车辆1至车辆3任一车辆采集的视频对应的交通场景整合得到的交通场景作为目标区域的交通场景，能够全面描述目标区域的交通状况。
步骤206，云端进行交通事故预测，判断所述目标区域的交通场景是否将成为交通事故场景。
在一个实施例中，参见图9示出的判断所述目标区域的交通场景是否将成为交通事故场景的一个可选的流程示意图，包括步骤2061至步骤2063，以下对各步骤进行说明。
步骤2061，提取交通事故模型库中各类型交通事故对应的交通事故场景的特征，所述交通事故模型库包括交通事故场景与交通事故类型的对应关系。
在一个实施例中，云端维护的交通事故模型库中包括不同类型交通事故的交通事故模型，交通事故模型包括交通事故场景与相应类型交通事 故的对应关系，不同交通事故的交通事故场景均包括道路场景、行人行为、车辆行为三种特征，当然，还可以包括其他类型的特征，如交通事故模型库的一个可选的示例如表3所示：
Figure PCTCN2017083762-appb-000002
表3
表3中示出的交通事故模型库仅仅是示例性的，在表3中，交通事故模型与交通事故场景是一一对应的关系。可选地，同一类型的交通事故可以对应多个交通事故场景。
步骤2062，将各类型交通事故模型对应的交通事故场景的特征与所述目标区域的交通场景的特征进行匹配，基于特征的匹配程度确定所述交通场景演变为相应交通事故场景的事故概率。
在一个实现方式中，各类型交通事故模型对应的交通事故场景的特征与所述目标区域的交通场景的各特征之间的匹配程度，决定了交通场景演变为相应交通事故场景的事故概率，因此，基于匹配程度与事故概率之间的正相关的数量关系，例如数量关系可以采用正比关系，或者采用任意具有单一上述趋势的曲线描述。
步骤2063，判定所述目标区域的交通场景将演变为事故概率最高的交通事故场景。
在一个实现方式中，对于以下情况，最高的事故概率低于事故概率阈值，这说明当前时刻目标区域的交通场景演变为交通事故场景的可能性很低，因此判定所述目标区域的交通场景将演变为事故概率最高、且事故概率高于事故概率阈值的交通事故场景，以确保交通事故预警的准确性。事故概率阈值可以根据后期的车辆驾驶者反馈的预警的准确率进行调整，如果车辆驾驶者反馈的预警的准确率过低，则说明事故概率阈值的取值较小影响了预警的精度，则调高(按照既定幅度)事故概率阈值，直至驾驶者反馈的预警的准确率达到实用标准。
步骤207，云端判定将成为交通事故场景时向所述目标区域中的车辆发出预警信息。
在一个实现方式中，判定目标区域的交通场景将演变成为交通事故场景时，向所述目标区域中所述交通事故场景的涉事车辆发出与交通事故场景对应类型的交通事故的预警，提示驾驶者注意驾驶以避免交通事故。
在另一个实现方式中，判定目标区域的交通场景将演变成为交通事故场景时，根据交通事故场景的类型分析出涉事车辆发生交通事故的原因，例如超速、并线等，并向涉事车辆发送行驶提示信息，指示所述涉事车辆按照提示行驶，例如请勿超速、请勿并线以防交通事故，以避免发生所述交通事故场景。
综上所述，本申请实施例中，通过车辆在区间中所处位置处采集的视频确定出交通场景，对区间中车辆所采集的视频进行合成以还原出区间的全景视频，从而借由全景视频确定出目标区间的车辆，基于目标区间的车辆所在位置的交通场景进行整合得到目标区间的完整的交通场景，从而可以利用交通事故模型对目标区间的完整的交通场景进行交通事故的预 判，在交通事故发生前对车辆进行预警以避免交通事故的发生，针对区间中车辆的实际交通场景进行交通事故的预测，针对性强且预测精度高，能够显著降低交通事故的发生率。
在还原区间的全景视频时，如果区间中的车流量较小，还原区间的全景视频完全依赖于区间中的车辆所采集的视频，会出现因采集方位单一、采集的视频较少而无法覆盖区间中全部区域，进而无法还原区间全部区域的全景视频的情况。
针对上述情况，在一个实施例中，还设置交通场景特征库，利用交通场景特征库中的视频作为利用移动终端采集视频合成全景视频的补充。示例性地，交通场景特征库可以包括以下至少一种：
1)利用车辆采集的视频提取的区间的静态交通场景
对于一个区间来说，从区间的每个车辆所采集到的视频中提取静态目标，也就是区间静止不变的目标，例如道路、立交桥，交通灯、车道线、道路牌和立杆等交通标识。基于所述静态目标合成所述区间的静态交通场景，利用所述静态交通场景迭代更新至交通场景特征库中对应所述区间的静态交通场景的视频，随着时间推移，区间的静态交通场景会不断完善，直至覆盖区间的全部区域。
2)利用第三方针对区间采集的监控视频提取的交通场景
从第三方，例如交通管理部门，的监控视频数据库获取的第三方针对所述区间中特定监控区域，例如十字路口、立交桥等，采集的监控视频，从监控视频中提取的静态目标，利用静态目标合成所述区间的静态交通场景，将所合成的静态交通场景迭代更新至所述交通场景特征库中。
参见图10，作为图6中示出的步骤204中云端得到区间的全景视频的一个可选的流程示意图，包括步骤2044和步骤2045，以下对各步骤进行说明。
步骤2044，从交通场景特征库获取静态交通场景。
在一个实施例中，当区间中的车流量低于车流量阈值时，即确定区间中车辆当前时刻采集的视频无法完整还原出区间的全景视频，从交通场景特征库获取静态交通场景。
步骤2045，将从所述交通场景特征库获取的所述区间的静态交通场景、连同所述区间中每个车辆从至少一个方位对所述区间进行采集得到的视频进行合成，对应得到所述区间的全景视频。
在一个实施例中，对所述区间中每个车辆采集得到的视频中各视频帧的目标特征、以及从交通场景特征库获取的视频中各视频帧的目标特征进行标记索引，以目标特征的描述的方式、或以样图显示的目标特征的方式对所述视频(包括移动终端在区间中所处位置利用摄像头采集的视频，以及从交通场景特征库获取的至少一种类型的视频)进行搜索得到具有相同目标特征的视频帧，对所搜索得到的具有相同目标特征的视频帧进行合成。
区间中的移动终端采集的视频主要驾驶者构建全景视频中的动态交通场景，从交通场景特征库获取的静态交通场景主要用于还原全景视频中的静止的景物，如建筑物等，二者相互补充从而能够完整地还原出区间的全景视频，进而能够准确确定目标区域的完整的交通场景，确保利用交通事故模型对目标区域的交通场景进行交通事故预判的准确性。
在一个实施例中，为了让交通事故模型库中的交通事故模型在交通事故的预测过程中进行自适应的更新，以能够对多样化的交通事故场景进行全面准确判断，参见图11示出的交通事故预警方法的一个可选的流程示意图，基于图6，还包括以下步骤：
步骤208，云端基于所述目标区域的交通场景、以及与所述事故概率最高的交通事故场景对应的交通事故类型的对应关系生成新的交通事故模 型。
步骤209，云端利用所述新的交通事故模型迭代更新所述交通事故模型库。
对于每次所预测出的新的交通事故场景，利用新的交通事故场景与对应的交通事故类型形成新的交通事故模型并以累加的方式更新至交通事故模型特征库，交通事故模型库中的交通事故模型跟随区间中的不断更新的交通事故场景实现了自适应学习，不需要人工对交通事故模型库进行手动更新，随着对区间中的交通事故的持续预测，利用交通事故模型库进行交通事故预测的精度会越来越高，确保了交通事故模型预测的准确度。
以下结合图12，对交通事故预警装置为移动终端时，对一个区间中的交通事故预警进行说明。对于多个区间中的交通事故预警可以参照对以下记载的技术方案而实施。与前述实施例不同的是，图12中示出的交通事故预警处理仅由移动终端协同完成，不需要云端的参与。
图12示出了根据本申请一实施例提供的交通事故预警方法的流程示意图，包括步骤301至步骤307，以下结合各步骤进行说明。
步骤301，区间中的车辆上的每个移动终端定位其所处车辆的位置，并从至少一个方位对车辆在区间中所处的位置进行采集得到视频。
如前所述，车辆的位置可以采用原始的地理坐标的方式或地理方格坐标的形式。可替换地，在一个实现方式中，不同区间的信息可以在移动终端中预先设置，移动终端之间彼此开放定位服务从而任一移动终端可以获取区间中其他移动终端的实时位置。
步骤302，区间中的每个移动终端将其所处的车辆的位置、以及采集得到视频发送至其他移动终端。
由于采集的视频后续需要发送至区间中其他的移动终端，考虑到不同移动终端的数据处理能力不同，移动终端可以将采集的视频进行压缩后 发送，以保证所有移动终端能够有足够空间接收视频并进行处理。
步骤303，区间中的移动终端解析所采集到的视频，得到所述区间中车辆所处位置的交通场景。
示例1)，在一个实现方式中，区间中每个车辆对所采集的视频以及区间中其他移动终端所发送的视频进行解析，得到区间中每个移动终端所处位置的交通场景。移动终端解析视频得到区间中所处位置的交通场景可以参照前述云端解析所采集到的视频得到区间中每个车辆所处位置的交通场景而实施。
示例2)在另一个实现方式中，区间中车辆进行能力协商确定出区间中处理能力最高的至少一个移动终端作为区间的节点，对区间中所有车辆采集的视频进行解析，得到区间中车辆所处位置的交通场景。
步骤304，区间中的每个移动终端合成所述区间中每个车辆从至少一个方位对所述区间进行采集得到的视频，对应得到所述区间的全景视频。
接续对前述示例1)进行说明，区间中每个移动终端基于所采集的视频以及所接收的视频进行全景合成，合成全景视频可以参照前述云端合成全景视频的记载而实施。
接续对前述示例2)进行说明，区间中作为节点的移动终端基于所采集的视频以及所接收的视频进行全景合成全景视频。
步骤305，区间中移动终端识别出位于所述全景视频中目标区域的车辆，基于所述目标区域中每个车辆所处区域的交通场景进行合成，得到所述目标区域的交通场景。
接续对前述示例1)进行说明，在一个实现方式中，区间中每个移动终端识别出位于所述全景视频中目标区域的车辆，基于所述目标区域中每个车辆所处区域的交通场景进行合成，得到所述目标区域的交通场景。同样地，识别出目标区域的交通场景可以参照前述云端识别目标区域的交通 场景的记载而实施。
接续对前述示例2)进行说明，在另一个实现方式中，区间中作为节点的移动终端识别出位于所述全景视频中目标区域的车辆，基于所述目标区域中每个车辆所处区域的交通场景进行合成，得到所述目标区域的交通场景。
步骤306，区间中移动终端判断所述目标区域的交通场景是否将成为交通事故场景。
接续对前述示例1)进行说明，在一个实现方式中，区间中每个移动终端判断所述目标区域的交通场景是否将成为交通事故场景。同样地，判断所述目标区域的交通场景是否将成为交通事故场景可以参照前述云端判断所述目标区域的交通场景是否将成为交通事故场景的记载而实施。
接续对前述示例2)进行说明，在另一个实现方式中，区间中作为节点的移动终端判断所述目标区域的交通场景是否将成为交通事故场景。
步骤307，区间中移动终端判定将成为交通事故场景时向所述目标区域中的车辆发出预警信息。
例如，对于任一移动终端，在判断出自身所处的区域的交通场景将成为交通事故场景时，移动终端发出预警并且向交通事故涉及的其他车辆中的移动终端发出预警。
接续对前述的交通事故预警装置的结构进行说明，参见图13示出的根据本申请一实施例的交通事故预警装置400的结构示意图，包括：获取模块401、解析模块402、合成模块403，识别模块404，预测模块405和预警模块406。该交通事故预警装置400还可包括交通事故模型库407和交通场景特征库408。以下对各模块进行说明。
获取模块401用于获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频。
在一个实现方式中，当交通事故预警装置400实现于移动终端时，获取模块401包括定位单元和采集单元。其中定位单元可以使用图3中示出的定位模块103实现，定位模块103接收GPS信号或北斗定位信号等方式对车辆进行定位，车辆的位置除了采用原始的地理坐标的方式，还可以采用地理方格的形式。采集单元用于在所述车辆所处位置从至少一个方位对所述区间进行采集得到视频。实际应用中，采集单元可以采用图3示出的摄像头106实现，采集单元可以实施为一个摄像头，或者，实施为多个摄像头106构成的模组以进行全方位的视频采集。另外，采集单元中还可以集成云台装置对摄像头106的采集方位进行调整。示例性地，采集单元设置在车辆的顶部，从而能够最大限度增大采集的视角。
解析模块402用于解析所采集到的视频，得到所述区间中每个车辆所处位置的交通场景。
在解析视频的得到交通场景的一些实施例中，解析模块402在每个车辆所采集到的视频中进行目标提取，得到以下目标至少之一：道路、车辆、行人和交通灯；对所提取的目标进行模式识别，得到以下特征至少之一：所述道路的道路场景、所述行人的行为类型、所述车辆的行为类型和所述交通灯的状态。
合成模块403合成所述区间中每个车辆从至少一个方位对所述区间进行采集得到的视频，对应得到所述区间的全景视频。
识别模块404识别出位于所述全景视频中目标区域的车辆，基于所述目标区域中每个车辆所处区域的交通场景进行合成，得到所述目标区域的交通场景。
在基于所述目标区域中每个车辆所处区域的交通场景进行合成的一些实施例中，合成模块403对所述区间中每个车辆采集得到的视频中各视频帧的目标特征进行标记索引；以目标特征的描述的方式、或以样图显示 的目标特征的方式对所述视频进行搜索得到具有相同目标特征的视频帧；对所搜索得到的具有相同目标特征的视频帧进行合成。
预测模块405判断所述目标区域的交通场景是否将成为交通事故场景。
在判断所述目标区域的交通场景是否将成为交通事故场景的一些实现方式中，所述预测模块405提取交通事故模型库407中各类型交通事故模型对应的交通事故场景的特征，所述交通事故模型库包括交通事故场景与交通事故类型的对应关系；将各类型交通事故对应的交通事故场景的特征与所述目标区域的交通场景的特征进行匹配，基于特征的匹配程度确定所述交通场景演变为相应交通事故场景的事故概率；判定所述目标区域的交通场景将演变为事故概率最高的交通事故场景。
预警模块406在所述预测模块405判定将成为交通事故场景时向所述目标区域中的车辆发出预警信息。
示例性地，预警模块406执行以下预警操作至少之一：向所述目标区域中所述交通事故场景涉及到的车辆发出与交通事故场景对应类型的交通事故的预警；向所述目标区域中所述交通事故场景涉及到的车辆发出行驶提示信息，用于指示所述涉事车辆按照提示行驶以避免所述交通事故场景。
需要指出的是，解析模块402、合成模块403、识别模块404、预测模块405和预警模块406是对交通事故预警装置在逻辑功能层面的划分，其中的任意模块可以合并实施，或者，其中的任意模块可以拆分多个模块的方式实施，解析模块402、合成模块403、识别模块404、预测模块405和预警模块406均可由图3中示出的处理器101通过执行存储器105存储的可执行指令实现。
在一个实施例中，为了保证合成的全景视频能够覆盖区间中全部区 域，解析模块402在每个车辆所采集到的视频中提取静态目标，基于所述静态目标合成所述区间的静态交通场景，利用所述静态交通场景迭代更新至交通场景特征库406中对应所述区间的静态交通场景；或者，从第三方监控视频数据库获取第三方针对所述区间中特定监控区域采集的监控视频，将所述监控视频更新至所述交通场景特征库406中对应所述区间的监控视频。相应地，合成模块还用于从交通场景特征库获取以下至少之一：第三方针对所述区间中特定监控区域的监控视频；对应所述区间中静态交通场景的视频；将从所述交通场景特征库获取的至少一种类型的视频、连同所述区间中每个车辆从至少一个方位对所述区间进行采集得到的视频，对应得到所述区间的全景视频。利用静态场景特征库中的静态场景的视频弥补车辆所采集视频中静态场景的不同，保证合成的全景视频中包括完整的交通场景。
在一个实施例中，为了让交通事故模型库中的交通事故模型在交通事故的预测过程中进行自适应的更新，以能够对多样化的交通事故场景进行全面准确判断，所述预警模块406，还用于基于所述目标区域的交通场景、以及与所述事故概率最高的交通事故场景对应的交通事故类型的对应关系生成新的交通事故模型；利用所述新的交通事故模型迭代更新所述交通事故模型库407。
再结合图14示出的一个实例进行说明，图14示出的区间中的目标区域的一个可选的示意图，以区间中的目标区域为十字路口为例。在该目标区域中，车辆B在十字路口的北段从北向南行驶，车辆A在十字路口的东段从东向西高速行驶，由于车辆A和车辆B的驾驶者视角的局限，车辆A和车辆B的驾驶者均无法感知到对方的存在，一旦车辆A闯红灯行驶，必然发生与车辆B相撞的事故。
基于本申请实施例提供的方案，车辆A和车辆B中设置的移动终端 对车辆所处位置进行视频采集以及定位，云端根据车辆A、车辆B中移动终端上传的视频以及位置合成十字路口的全景视频，在合成全景视频时还可以结合交通场景特征库对应该十字路口的视频进行合成，从而还原出当前时刻的十字路口完整的交通场景的全景视频。
云端对全景视频解析出十字路口的交通场景，如表4所示：
  道路场景 车辆行为 行人行为
车辆B **北路、北向畅通 穿越十字路口 无行人
车辆A **东路，东西向畅通 闯红灯行驶 无行人
表4
利用交通事故模型库中的交通事故模型并基于全景视频进行交通事故预测，匹配到两车在十字路口相撞的交通事故模型，如表5所示：
Figure PCTCN2017083762-appb-000003
表5
预测到车辆A由于闯红灯将导致与车辆B相撞的事故，实时向车辆A和车辆B发出预警，以避免事故发生。例如，移动终端可以在显示屏上显示提示信息，并发出蜂鸣报警提示，如图15所示，对于车辆A的驾驶者提示闯红灯即将撞车，对于车辆B的驾驶者提示有车辆在另一方向闯红灯行驶注意规避。
本申请一个实施例中还提供一种计算机存储介质，其中存储有可执 行指令，所述可执行指令用于执行前述的交通事故预警方法。
综上所述，本申请实施例中通过采集车辆在区间中所处位置的实时视频，还原出区间的全景视频，基于全景视频能够精确确定目标区域的实时交通场景，从而利用交通事故模型对目标区间的交通场景进行交通事故的预判，这样在交通事故发生前对相关车辆进行预警以避免交通事故的发生，由于是针对区间中车辆的实际交通场景(也就是行车状况)进行交通事故的预测，针对性强且准确率高，能够显著降低交通事故的发生率。
本申请实施例所述集成的模块如果以软件功能模块的形式实现并作为独立的产品销售或使用时，也可以存储在一个计算机可读取存储介质中。基于这样的理解，本领域内的技术人员应明白，本申请的实施例可提供为方法、系统、或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质上实施的计算机程序产品的形式，所述存储介质包括但不限于U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁盘存储器、CD-ROM、光学存储器等。
本申请是根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处 理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
尽管已描述了本申请的实施例，但本领域内的技术人员一旦得知了基本创造性概念，则可对这些实施例做出另外的变更和修改。所以，所附权利要求意欲解释为包括实施例以及落入本申请范围的所有变更和修改。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (17)

  1. 一种交通事故预警方法，其特征在于，所述方法包括：
    获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频；
    解析所采集到的视频，得到每个所述车辆所处位置的交通场景；
    合成在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频；
    识别出位于所述全景视频中的目标区域中的车辆，对所述目标区域中的车辆的车辆所处位置的交通场景进行合成，得到所述目标区域的交通场景；
    基于所述目标区域的交通场景对所述目标区域进行交通事故预测；以及
    当判断所述目标区域的交通场景将成为交通事故场景时，向所述目标区域中的至少一个车辆发出预警信息。
  2. 如权利要求1所述的方法，其特征在于，所述解析所采集到的视频，得到所述区间中每个车辆所处位置的交通场景，包括：
    从在每个所述车辆所处位置处采集得到的视频中进行目标提取，得到以下目标至少之一：道路、车辆、行人和交通标识；
    对所提取的目标进行模式识别，得到以下特征至少之一：所述道路的道路场景、所述行人的行为类型、所述车辆的行为类型和所述交通标识的状态。
  3. 如权利要求1所述的方法，其特征在于，所述合成在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频，包括：
    从交通场景特征库获取所述区间的静态交通场景；
    将所获取的所述区间的静态交通场景、以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频进行合成，得到所述区间的全景视频。
  4. 如权利要求3所述的方法，其特征在于，所述方法还包括：
    从在每个所述车辆所处位置处采集得到的视频中提取静态目标，基于所述静态目标合成所述区间的静态交通场景，将所述静态交通场景迭代更新至所述交通场景特征库中；
    和/或，
    从第三方监控视频数据库获取监控视频，基于从所述监控视频中提取的静态目标合成所述区间的静态交通场景，将所合成的静态交通场景迭代更新至所述交通场景特征库中，其中，所述监控视频为第三方针对所述区间中特定监控区域采集的视频。
  5. 如权利要求1所述的方法，其特征在于，所述合成在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频，包括：
    对每个所述车辆所处位置处采集得到的视频中每个视频帧的目标特征进行标记索引；
    以目标特征的描述的方式、或以样图显示的目标特征的方式对所述视 频进行搜索得到具有相同目标特征的视频帧；
    对所述具有相同目标特征的视频帧进行合成。
  6. 如权利要求1所述的方法，其特征在于，所述基于所述目标区域的交通场景对所述目标区域进行交通事故预测，包括：
    提取所述目标区域的交通场景的特征；
    将交通事故模型库中各类型交通事故对应的交通事故场景的特征与所述目标区域的交通场景的特征进行匹配，基于特征的匹配程度确定所述目标区域的交通场景演变为交通事故的概率，其中所述交通事故模型库包括交通事故场景与交通事故类型的对应关系；
    判定所述目标区域的交通场景将演变为事故概率最高的交通事故场景。
  7. 如权利要求6所述的方法，其特征在于，所述方法还包括：
    基于所述目标区域的交通场景、以及与所述目标区域的交通场景对应的交通事故类型生成新的交通事故模型；
    利用所述新的交通事故模型迭代更新所述交通事故模型库。
  8. 如权利要求1所述的方法，其特征在于，
    当判断所述目标区域的交通场景将成为交通事故场景时，向所述目标区域中的至少一个车辆发出预警信息，包括执行以下预警操作至少之一：
    向所述目标区域中所述交通事故场景涉及到的车辆发出与所述交通事故场景对应类型的交通事故预警；
    向所述目标区域中所述交通事故场景涉及到的车辆发出行驶提示信息，用于指示所述车辆按照提示行驶以避免所述交通事故场景。
  9. 一种交通事故预警装置，其特征在于，包括：
    获取模块，用于获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频；
    解析模块，用于解析所采集到的视频，得到每个所述车辆所处位置的交通场景；
    合成模块，用于合成在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频；
    识别模块，用于识别出位于所述全景视频中的目标区域中的车辆，对所述目标区域中的车辆的车辆所处位置的交通场景进行合成，得到所述目标区域的交通场景；
    预测模块，用于基于所述目标区域的交通场景对所述目标区域进行交通事故预测；以及
    预警模块，用于当判断所述目标区域的交通场景将成为交通事故场景时，向所述目标区域中的至少一个车辆发出预警信息。
  10. 如权利要求9所述的交通事故预警装置，其特征在于，
    所述解析模块，还用于从在每个所述车辆所处位置处采集得到的视频中进行目标提取，得到以下目标至少之一：道路、车辆、行人和交通标识；
    所述解析模块，还用于对所提取的目标进行模式识别，得到以下特征至少之一：所述道路的道路场景、所述行人的行为类型、所述车辆的行为类型和所述交通标识的状态。
  11. 如权利要求9所述的交通事故预警装置，其特征在于，
    所述合成模块，还用于从交通场景特征库获取所述区间的静态交通场 景；以及
    将所获取的所述区间的静态交通场景、以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频进行合成，得到所述区间的全景视频。
  12. 如权利要求11所述的交通事故预警装置，其特征在于，
    所述解析模块，还用于从在每个所述车辆所处位置处采集得到的视频中提取静态目标，基于所述静态目标合成所述区间的静态交通场景，将所述静态交通场景迭代更新至所述交通场景特征库中；
    或者，从第三方监控视频数据库获取监控视频，基于从所述监控视频中提取的静态目标合成所述区间的静态交通场景，将所合成的静态交通场景迭代更新至所述交通场景特征库中，其中，所述监控视频为第三方针对所述区间中特定监控区域采集的视频。
  13. 如权利要求9所述的交通事故预警装置，其特征在于，
    所述合成模块，还用于对每个所述车辆所处位置处采集得到的视频中每个视频帧的目标特征进行标记索引；
    以目标特征的描述的方式、或以样图显示的目标特征的方式对所述视频进行搜索得到具有相同目标特征的视频帧；以及
    对所述具有相同目标特征的视频帧进行合成。
  14. 如权利要求9所述的交通事故预警装置，其特征在于，
    所述预测模块，还用于提取所述目标区域的交通场景的特征；
    将交通事故模型库中各类型交通事故对应的交通事故场景的特征与所述目标区域的交通场景的特征进行匹配，基于特征的匹配程度确定所述目 标区域的交通场景演变为交通事故的概率，其中所述交通事故模型库包括交通事故场景与交通事故类型的对应关系；以及
    判定所述目标区域的交通场景将演变为事故概率最高的交通事故场景。
  15. 如权利要求14所述的交通事故预警装置，其特征在于，
    所述预测模块，还用于基于所述目标区域的交通场景、以及与所述目标区域的交通场景对应的交通事故类型生成新的交通事故模型；利用所述新的交通事故模型迭代更新所述交通事故模型库。
  16. 如权利要求9所述的交通事故预警装置，其特征在于，
    所述预警模块，还用于执行以下预警操作至少之一：
    向所述目标区域中所述交通事故场景涉及到的车辆发出与交通事故场景对应类型的交通事故预警；
    向所述目标区域中所述交通事故场景涉及到的车辆发出行驶提示信息，用于指示所述车辆按照提示行驶以避免所述交通事故场景。
  17. 一种交通事故预警装置，其特征在于，所述装置包括处理器和存储器，所述处理器执行存储器中存储的程序以执行：
    获取区间中多个车辆的位置以及在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频；
    解析所采集到的视频，得到每个所述车辆所处位置的交通场景；
    合成在每个所述车辆所处位置处从至少一个方位对所述区间进行采集得到的视频，得到所述区间的全景视频；
    识别出位于所述全景视频中的目标区域中的车辆，对所述目标区域中 的车辆的车辆所处位置的交通场景进行合成，得到所述目标区域的交通场景；
    基于所述目标区域的交通场景对所述目标区域进行交通事故预测；以及
    当判断所述目标区域的交通场景将成为交通事故场景时，向所述目标区域中的至少一个车辆发出预警信息。
PCT/CN2017/083762 2016-05-10 2017-05-10 交通事故预警方法及交通事故预警装置 Ceased WO2017193933A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP17795551.5A EP3457380A4 (en) 2016-05-10 2017-05-10 METHOD FOR TRAFFIC PRECAUTION AND DEVICE FOR TRAFFIC PRECAUTION
US16/046,716 US10403138B2 (en) 2016-05-10 2018-07-26 Traffic accident warning method and traffic accident warning apparatus

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610308302.4A CN105761500B (zh) 2016-05-10 2016-05-10 交通事故处理方法及交通事故处理装置
CN201610308302.4 2016-05-10

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/046,716 Continuation-In-Part US10403138B2 (en) 2016-05-10 2018-07-26 Traffic accident warning method and traffic accident warning apparatus

Publications (1)

Publication Number Publication Date
WO2017193933A1 true WO2017193933A1 (zh) 2017-11-16

Family

ID=56323898

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/083762 Ceased WO2017193933A1 (zh) 2016-05-10 2017-05-10 交通事故预警方法及交通事故预警装置

Country Status (4)

Country Link
US (1) US10403138B2 (zh)
EP (1) EP3457380A4 (zh)
CN (1) CN105761500B (zh)
WO (1) WO2017193933A1 (zh)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110490126A (zh) * 2019-08-15 2019-11-22 成都睿晓科技有限公司 一种基于人工智能的保险柜安全管控系统
WO2020000834A1 (zh) * 2018-06-28 2020-01-02 平安科技（深圳）有限公司 交通事故处理方法、系统及服务器
CN110852302A (zh) * 2019-11-20 2020-02-28 北京明略软件系统有限公司 事故嫌疑身份信息确定方法及相关设备
CN111684505A (zh) * 2018-09-13 2020-09-18 Jvc建伍株式会社 影像记录控制装置、影像记录系统、影像记录方法以及影像记录控制程序
CN113128294A (zh) * 2019-12-31 2021-07-16 深圳云天励飞技术有限公司 一种道路事件取证方法、装置、电子设备及存储介质
CN113380034A (zh) * 2021-06-10 2021-09-10 智道网联科技(北京)有限公司 事故定位方法和装置、电子设备和计算机可读存储介质
CN114462695A (zh) * 2022-01-27 2022-05-10 东南大学 一种基于道路安全模型的车辆安全曝光量分析方法
CN117455209A (zh) * 2023-12-26 2024-01-26 山东智慧云通网络科技有限公司 一种基于地理位置的智慧交警管理服务系统及方法

Families Citing this family (71)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105761500B (zh) 2016-05-10 2019-02-22 腾讯科技（深圳）有限公司 交通事故处理方法及交通事故处理装置
CN106228493A (zh) * 2016-08-01 2016-12-14 深圳市永兴元科技有限公司 车辆事故处理方法及装置
CN106781436B (zh) * 2016-12-19 2020-03-06 东软集团股份有限公司 交通事故处理方法及装置
US20200045298A1 (en) * 2017-02-27 2020-02-06 Thomson Licensing Method, system and apparatus for visual effects
CN107066965A (zh) * 2017-04-11 2017-08-18 北京汽车集团有限公司 检测交通标识的方法及装置
US10606767B2 (en) * 2017-05-19 2020-03-31 Samsung Electronics Co., Ltd. Ethernet-attached SSD for automotive applications
CN107507445A (zh) * 2017-08-17 2017-12-22 千寻位置网络有限公司 基于高精度定位自动播报交通事故和拥堵车道的方法
CN107633570A (zh) * 2017-08-28 2018-01-26 武汉六点整北斗科技有限公司 一种交通事故的快撤实现方法及相关产品
CN109935107B (zh) * 2017-12-18 2023-07-14 姜鹏飞 一种提升交通通视范围的方法及装置
US10607484B2 (en) * 2017-12-28 2020-03-31 Intel Corporation Privacy-preserving distributed visual data processing
US11349753B2 (en) * 2017-12-28 2022-05-31 Intel Corporation Converged routing for distributed computing systems
US10797863B2 (en) 2017-12-28 2020-10-06 Intel Corporation Multi-domain cascade convolutional neural network
CN111699679B (zh) * 2018-04-27 2023-08-01 上海趋视信息科技有限公司 交通系统监控和方法
CN108986468A (zh) * 2018-08-01 2018-12-11 平安科技（深圳）有限公司 交通事故的处理方法、装置、计算机设备及计算机存储介质
US11138418B2 (en) * 2018-08-06 2021-10-05 Gal Zuckerman Systems and methods for tracking persons by utilizing imagery data captured by on-road vehicles
CN109147340A (zh) * 2018-08-30 2019-01-04 平安科技（深圳）有限公司 违法停车案件的鉴别方法、装置和计算机设备
CN110969858B (zh) * 2018-09-29 2022-05-13 比亚迪股份有限公司 车祸处理的方法、装置和存储介质及电子设备
CN109543588A (zh) * 2018-11-16 2019-03-29 广州慧睿思通信息科技有限公司 交通事故责任判定的方法、装置、系统、服务平台及介质
CN109661692B (zh) * 2018-11-22 2021-07-30 深圳市锐明技术股份有限公司 交通事件预测方法、装置及终端设备
CN109544725B (zh) * 2018-12-03 2021-01-12 浪潮集团有限公司 一种基于事件驱动的自动驾驶事故智能处理方法
CN109413415B (zh) * 2018-12-12 2024-03-29 清华大学苏州汽车研究院（吴江） 一种摄像头控制器测试系统及测试方法
CN110491139B (zh) * 2019-07-11 2021-02-19 深圳市轱辘车联数据技术有限公司 一种交通事故责任预警管理方法及相关装置
CN114283040A (zh) * 2019-12-04 2022-04-05 蚂蚁胜信(上海)信息技术有限公司 交通事故处理方法和装置、电子设备
CN111369807B (zh) * 2020-03-24 2022-04-12 北京百度网讯科技有限公司 一种交通事故的检测方法、装置、设备和介质
US11904854B2 (en) 2020-03-30 2024-02-20 Toyota Research Institute, Inc. Systems and methods for modeling pedestrian activity
CN113140107B (zh) * 2020-03-30 2022-09-16 北京高德云图科技有限公司 交通信息处理方法、装置、电子设备及计算机存储介质
CN111563425B (zh) * 2020-04-22 2023-04-07 蘑菇车联信息科技有限公司 交通事件识别方法及电子设备
CN111542019A (zh) * 2020-04-23 2020-08-14 浙江省交通规划设计研究院有限公司 交通事故告警方法、装置、路侧设备和可读存储介质
CN111526337B (zh) * 2020-05-08 2021-12-17 三一重机有限公司 一种工程机械的预警系统、预警方法及工程机械
CN113685179B (zh) * 2020-05-19 2024-11-22 郑州煤机智能工作面科技有限公司 一种煤矿综采工作面自动化控制系统及其控制方法
CN113808386B (zh) * 2020-06-17 2025-11-11 中兴通讯股份有限公司 交通事故处理方法及装置
CN111814625B (zh) * 2020-06-29 2024-04-19 广东荣文科技集团有限公司 智能交通事故处理方法及相关装置
CN111739344B (zh) * 2020-06-29 2022-11-08 北京百度网讯科技有限公司 一种预警方法、装置及电子设备
CN114079865B (zh) * 2020-08-10 2024-12-27 罗伯特·博世有限公司 移动目标标识方法与节点
WO2022066026A1 (en) * 2020-09-28 2022-03-31 Motorola Solutions, Inc Device, method and system for providing a notification of a distinguishing activity
CN112073696B (zh) * 2020-09-30 2022-08-23 富盛科技股份有限公司 一种基于无人机的现场交通事故采集执法系统及方法
CN112509315B (zh) * 2020-11-04 2022-02-15 杭州远眺科技有限公司 一种基于视频分析的交通事故检测方法
CN112560609B (zh) 2020-12-03 2022-11-15 北京百度网讯科技有限公司 路况预估方法、建立路况预估模型的方法及对应装置
CN114612876B (zh) * 2020-12-09 2025-09-19 阿里巴巴集团控股有限公司 获取交通事件的方法和导航终端
CN112735124B (zh) * 2020-12-16 2022-05-20 阿波罗智联(北京)科技有限公司 交通数据的分析方法、装置、设备、车辆及存储介质
CN113077634A (zh) * 2021-03-19 2021-07-06 上海电机学院 一种辅助交通监控的方法
CN112949598B (zh) * 2021-04-07 2022-02-11 吴洲 一种公路建筑施工安全智能识别系统
CN113033471A (zh) * 2021-04-15 2021-06-25 北京百度网讯科技有限公司 交通异常检测方法、装置、设备、存储介质以及程序产品
CN115331429A (zh) * 2021-04-20 2022-11-11 孙素萍 基于智慧城市的交通风险感知系统
CN113256969B (zh) * 2021-04-30 2022-08-16 山东金宇信息科技集团有限公司 一种高速公路的交通事故预警方法、设备及介质
US11741721B2 (en) * 2021-05-05 2023-08-29 Gm Cruise Holdings Llc Automatic detection of roadway signage
CN113223292A (zh) * 2021-05-06 2021-08-06 朱晓东 一种基于北斗交通事故处理的系统
CN115346395A (zh) * 2021-05-10 2022-11-15 北京万集科技股份有限公司 控制停车场门禁的方法和装置
CN113538904A (zh) * 2021-06-22 2021-10-22 浙江工业大学 一种基于时空前序编解码预测器的交通事故识别方法
CN113313075B (zh) * 2021-06-29 2024-02-02 杭州海康威视系统技术有限公司 目标对象位置关系分析方法、装置、存储介质及电子设备
CN113657299B (zh) * 2021-08-20 2024-10-15 青岛海信网络科技股份有限公司 交通事故确定方法及电子设备
CN113823084B (zh) * 2021-08-20 2023-10-27 广西北投交通养护科技集团有限公司 一种交通引导方法、装置、设备、系统及可读存储介质
CN113743331B (zh) * 2021-09-08 2024-02-23 济南博观智能科技有限公司 一种卸油运作监控方法、装置、设备及存储介质
JP7552567B2 (ja) * 2021-12-10 2024-09-18 トヨタ自動車株式会社 監視映像出力システム、監視映像出力方法
CN117037460A (zh) * 2022-04-28 2023-11-10 浙江宇视科技有限公司 一种交通设施调整方法、设备和装置
IT202200012116A1 (it) * 2022-06-08 2023-12-08 Fiat Ricerche "Autoveicolo comprendente un sistema di assistenza alla guida in comunicazione con un sistema informatico remoto, apparato e procedimento di assistenza alla guida di un autoveicolo corrispondenti"
CN115171025B (zh) * 2022-07-26 2026-03-13 智道网联科技(北京)有限公司 存储视频的方法、装置和电子设备
JP7606651B2 (ja) * 2022-08-25 2024-12-25 株式会社 ミックウェア 情報処理装置、移動端末、情報処理方法、およびプログラム
CN115641217B (zh) * 2022-10-18 2023-08-15 深圳市杰讯互联科技有限公司 一种用于交通事故快处快赔的系统
US12534097B1 (en) * 2022-11-01 2026-01-27 Samsara Inc. Driver alerting and feedback
CN116168531A (zh) * 2022-11-29 2023-05-26 上海腾盛智能安全科技股份有限公司 基于雷达与视频的隧道安全管理方法、装置、设备及介质
CN115935642B (zh) * 2022-12-05 2025-08-01 长沙理工大学 基于事故信息的智能车极端测试场景自动生成方法与系统
CN116612443A (zh) * 2023-04-26 2023-08-18 合肥工业大学 一种基于多源移动视频融合的交通场景还原方法及系统
CN116597688A (zh) * 2023-05-10 2023-08-15 智道网联科技(北京)有限公司 隧道场景的预警方法、装置及电子设备、存储介质
CN116469254B (zh) * 2023-06-15 2023-09-08 浙江口碑网络技术有限公司 一种信息处理方法和装置
CN117351696A (zh) * 2023-08-31 2024-01-05 智慧互通科技股份有限公司 基于场景解析的交通智能管控方法及系统
CN117216717A (zh) * 2023-09-26 2023-12-12 一汽解放汽车有限公司 驾驶场景预测方法、装置、计算机设备、存储介质和产品
CN118038334B (zh) * 2024-03-15 2024-09-13 中关村科学城城市大脑股份有限公司 事故处理方法、装置、电子设备和计算机可读介质
CN118644984B (zh) * 2024-06-24 2025-07-01 浙江安程市政交通设施有限公司 基于智能警示牌的交通事故告警系统及方法
WO2026090901A1 (zh) * 2024-10-30 2026-05-07 富士通株式会社 检测事故的方法和装置、以及存储介质
CN119832128A (zh) * 2025-03-13 2025-04-15 交通运输部规划研究院 一种水上交通事故情景推演方法及系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102157069A (zh) * 2011-04-01 2011-08-17 北京航空航天大学 一种基于视频融合的平面交叉口交通冲突检测方法
CN103236158A (zh) * 2013-03-26 2013-08-07 中国公路工程咨询集团有限公司 一种基于视频的交通事故实时预警方法
CN103985254A (zh) * 2014-05-29 2014-08-13 四川川大智胜软件股份有限公司 一种用于大场景交通监视的多视点视频融合及交通参数采集方法
US20140300690A1 (en) * 2002-10-11 2014-10-09 Brian KwangShik Hong Peripheral Viewing System for a Vehicle
CN105407278A (zh) * 2015-11-10 2016-03-16 北京天睿空间科技股份有限公司 一种全景视频交通态势监控系统及方法
CN105761500A (zh) * 2016-05-10 2016-07-13 腾讯科技（深圳）有限公司 交通事故处理方法及交通事故处理装置

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070069920A1 (en) * 2005-09-23 2007-03-29 A-Hamid Hakki System and method for traffic related information display, traffic surveillance and control
JP4893945B2 (ja) * 2007-02-06 2012-03-07 株式会社デンソー 車両周辺監視装置
US9183744B2 (en) * 2008-01-29 2015-11-10 Here Global B.V. Method for providing images of traffic incidents
US8345098B2 (en) * 2008-03-17 2013-01-01 International Business Machines Corporation Displayed view modification in a vehicle-to-vehicle network
CN101751782A (zh) * 2009-12-30 2010-06-23 北京大学深圳研究生院 一种基于多源信息融合的十字路口交通事件自动检测系统
JP2011198261A (ja) * 2010-03-23 2011-10-06 Hiromitsu Hama 対象物認識システム及び該システムを利用する監視システム、見守りシステム
EP2442291B1 (en) * 2010-10-13 2013-04-24 Harman Becker Automotive Systems GmbH Traffic event monitoring
US8532914B2 (en) * 2011-11-11 2013-09-10 Verizon Patent Licensing Inc. Live traffic congestion detection
CN102496000B (zh) * 2011-11-14 2013-05-08 电子科技大学 一种城市交通事故检测方法
JP6127659B2 (ja) * 2013-03-29 2017-05-17 富士通株式会社 運転支援装置及び運転支援方法
TWI552897B (zh) * 2013-05-17 2016-10-11 財團法人工業技術研究院 影像動態融合方法與裝置
EP2827259A1 (en) * 2013-07-16 2015-01-21 Alcatel Lucent Method for providing image data for a location
CN103794087A (zh) * 2014-02-17 2014-05-14 东南大学 基于无线协同的运动物体穿行道路辅助避撞方法和系统
WO2016030934A1 (ja) * 2014-08-25 2016-03-03 パイオニア株式会社 情報処理装置、制御方法、プログラム及び記憶媒体
CN105096659B (zh) * 2015-09-30 2018-06-15 宇龙计算机通信科技(深圳)有限公司 一种交通预警方法、装置及系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140300690A1 (en) * 2002-10-11 2014-10-09 Brian KwangShik Hong Peripheral Viewing System for a Vehicle
CN102157069A (zh) * 2011-04-01 2011-08-17 北京航空航天大学 一种基于视频融合的平面交叉口交通冲突检测方法
CN103236158A (zh) * 2013-03-26 2013-08-07 中国公路工程咨询集团有限公司 一种基于视频的交通事故实时预警方法
CN103985254A (zh) * 2014-05-29 2014-08-13 四川川大智胜软件股份有限公司 一种用于大场景交通监视的多视点视频融合及交通参数采集方法
CN105407278A (zh) * 2015-11-10 2016-03-16 北京天睿空间科技股份有限公司 一种全景视频交通态势监控系统及方法
CN105761500A (zh) * 2016-05-10 2016-07-13 腾讯科技（深圳）有限公司 交通事故处理方法及交通事故处理装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3457380A4 *

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020000834A1 (zh) * 2018-06-28 2020-01-02 平安科技（深圳）有限公司 交通事故处理方法、系统及服务器
CN111684505B (zh) * 2018-09-13 2022-10-18 Jvc建伍株式会社 影像记录控制装置、影像记录系统、影像记录方法以及影像记录控制程序
CN111684505A (zh) * 2018-09-13 2020-09-18 Jvc建伍株式会社 影像记录控制装置、影像记录系统、影像记录方法以及影像记录控制程序
EP3754625A4 (en) * 2018-09-13 2021-05-19 JVCKenwood Corporation Video recording control device, video recording system, video recording method, and video recording control program
US11321574B2 (en) 2018-09-13 2022-05-03 Jvckenwood Corporation Video recording control device, video recording system, video recording method, and non-transitory storage medium
CN110490126A (zh) * 2019-08-15 2019-11-22 成都睿晓科技有限公司 一种基于人工智能的保险柜安全管控系统
CN110490126B (zh) * 2019-08-15 2023-04-18 成都睿晓科技有限公司 一种基于人工智能的保险柜安全管控系统
CN110852302A (zh) * 2019-11-20 2020-02-28 北京明略软件系统有限公司 事故嫌疑身份信息确定方法及相关设备
CN113128294B (zh) * 2019-12-31 2024-06-11 深圳云天励飞技术有限公司 一种道路事件取证方法、装置、电子设备及存储介质
CN113128294A (zh) * 2019-12-31 2021-07-16 深圳云天励飞技术有限公司 一种道路事件取证方法、装置、电子设备及存储介质
CN113380034A (zh) * 2021-06-10 2021-09-10 智道网联科技(北京)有限公司 事故定位方法和装置、电子设备和计算机可读存储介质
CN114462695A (zh) * 2022-01-27 2022-05-10 东南大学 一种基于道路安全模型的车辆安全曝光量分析方法
CN114462695B (zh) * 2022-01-27 2024-05-03 东南大学 一种基于道路安全模型的车辆安全曝光量分析方法
CN117455209B (zh) * 2023-12-26 2024-03-12 山东智慧云通网络科技有限公司 一种基于地理位置的智慧交警管理服务系统及方法
CN117455209A (zh) * 2023-12-26 2024-01-26 山东智慧云通网络科技有限公司 一种基于地理位置的智慧交警管理服务系统及方法

Also Published As

Publication number Publication date
EP3457380A4 (en) 2020-04-15
CN105761500A (zh) 2016-07-13
US20180330610A1 (en) 2018-11-15
EP3457380A1 (en) 2019-03-20
US10403138B2 (en) 2019-09-03
CN105761500B (zh) 2019-02-22

Similar Documents

Publication Publication Date Title
US10403138B2 (en) Traffic accident warning method and traffic accident warning apparatus
US11967109B2 (en) Vehicle localization using cameras
CN110174093B (zh) 定位方法、装置、设备和计算机可读存储介质
JP6494719B2 (ja) 交通信号マップ作成及び検出
EP3660737B1 (en) Method, apparatus, and system for providing image labeling for cross view alignment
EP3644013B1 (en) Method, apparatus, and system for location correction based on feature point correspondence
WO2018083999A1 (ja) 表示方法及び表示装置
JP2022529337A (ja) 交通道路用のデジタル復元方法、装置及びシステム
US12097856B2 (en) System and method for adjusting a yielding space of a platoon
JP7238821B2 (ja) 地図生成システム及び地図生成プログラム
CN115294544A (zh) 驾驶场景分类方法、装置、设备及存储介质
US10949707B2 (en) Method, apparatus, and system for generating feature correspondence from camera geometry
CN103890784A (zh) 用于在传感器数据的流中探测对象的设备和方法
CN113902047B (zh) 图像元素匹配方法、装置、设备以及存储介质
US11410432B2 (en) Methods and systems for displaying animal encounter warnings in vehicles
CN108986463B (zh) 一种路况信息处理方法、装置及电子设备
CN116958915B (zh) 目标检测方法、装置、电子设备及存储介质
HK40098998A (zh) 车辆控制方法、装置、设备及存储介质

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17795551

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2017795551

Country of ref document: EP

Effective date: 20181210

WWW Wipo information: withdrawn in national office

Ref document number: 2017795551

Country of ref document: EP