WO2017041390A1 - 高精地图数据的处理方法、装置、存储介质和设备 - Google Patents

高精地图数据的处理方法、装置、存储介质和设备 Download PDF

Info

Publication number
WO2017041390A1
WO2017041390A1 PCT/CN2015/099012 CN2015099012W WO2017041390A1 WO 2017041390 A1 WO2017041390 A1 WO 2017041390A1 CN 2015099012 W CN2015099012 W CN 2015099012W WO 2017041390 A1 WO2017041390 A1 WO 2017041390A1
Authority
WO
WIPO (PCT)
Prior art keywords
map data
original map
point cloud
data
original
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/CN2015/099012
Other languages
English (en)
French (fr)
Inventor
蒋昭炎
关书伟
宋良
晏阳
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.)
Baidu Online Network Technology Beijing Co Ltd
Original Assignee
Baidu Online Network Technology Beijing 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 Baidu Online Network Technology Beijing Co Ltd filed Critical Baidu Online Network Technology Beijing Co Ltd
Priority to KR1020187007384A priority Critical patent/KR102037820B1/ko
Priority to JP2018512859A priority patent/JP6664470B2/ja
Priority to US15/758,623 priority patent/US10628712B2/en
Priority to EP15903493.3A priority patent/EP3343503B1/en
Publication of WO2017041390A1 publication Critical patent/WO2017041390A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three-dimensional [3D] modelling for computer graphics
    • G06T17/05—Geographic models
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/28—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
    • G01C21/30—Map- or contour-matching
    • G01C21/32—Structuring or formatting of map data
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/38—Registration of image sequences
    • G—PHYSICS
    • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09B—EDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
    • G09B29/00—Maps; Plans; Charts; Diagrams, e.g. route diagram
    • G09B29/003—Maps
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10028—Range image; Depth image; 3D point clouds
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20212—Image combination
    • G06T2207/20221—Image fusion; Image merging
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2210/00—Indexing scheme for image generation or computer graphics
    • G06T2210/56—Particle system, point based geometry or rendering

Definitions

  • Embodiments of the present invention relate to a map data processing technology, and in particular, to a method, an apparatus, a storage medium, and a device for processing high-precision map data.
  • the acquisition of the map image data generally utilizes an Intertial Measurement Uint (IMU), a Global Positioning System (GPS), or a Light Detection And Ranging (LIDAR).
  • IMU Intertial Measurement Uint
  • GPS Global Positioning System
  • LIDAR Light Detection And Ranging
  • the IMU mainly includes two parts, a gyroscope and an accelerometer, which directly affect the accuracy of the measured image data of the IMU.
  • the gyroscope mainly measures the angular velocity information of the moving object
  • the accelerometer mainly measures the acceleration information of the moving object
  • the IMU calculates the motion form of the moving object according to the measured information.
  • GPS mainly includes three parts: space constellation, ground monitoring and user equipment. These three parts directly affect the accuracy of image data obtained by GPS.
  • the space constellation is generally composed of 24 satellites, which is used for real-time measurement of image data of ground objects, and ground monitoring data of the object measured by the satellite transmitted by the satellite receiving the spatial constellation, and the user equipment calculates the map data of the object. And get the position information of the object and so on.
  • LIDAR mainly uses IMU and GPS for laser scanning, and the measured data is point cloud data, ie The measured data is a series of discrete points of the surface model of the object, and the point cloud data includes spatial three-dimensional information of the object and laser intensity information.
  • the sensor in LIDAR emits a laser beam to the ground or the surface of the object. The laser beam is emitted by an obstacle. The energy reflected by the laser beam is recorded by the sensor and used as laser intensity information, and the laser beam information can be used to calculate the laser beam. Reflectivity.
  • the above-mentioned tools are used to acquire the acquisition trajectory of the map data multiple times. After the map data collection is completed, the GPS or IMU is used to correlate the multiple corresponding map data obtained, and finally the high-precision map is obtained.
  • the prior art has the following drawbacks: when using the GPS or IMU to correlate the multiple corresponding map data obtained by the acquisition, due to the limitation of the measurement and calculation accuracy of the tool itself, the objects in the map image obtained after the association appear. Misplacement causes the object in the map image that the user sees to have a ghost image, which seriously affects the user experience.
  • the main purpose of the embodiments of the present invention is to provide a method, an apparatus, a storage medium, and a device for processing high-precision map data.
  • the problem of misalignment and ghosting of objects in a map image that occurs when correlating image data acquired multiple times is solved.
  • an embodiment of the present invention provides a method for processing high-precision map data, where the method includes:
  • Each of the original map data is subjected to fusion processing by using an object matching the feature of the object as a reference.
  • the embodiment of the present invention further provides a processing device for high-precision map data, where the device includes:
  • An acquisition module configured to perform at least two acquisitions of the map data along the target trajectory to obtain at least two original map data respectively;
  • An identification module configured to separately identify an object in the original map data, and extract an object feature
  • a matching module configured to match object features in each of the original map data
  • the fusion processing module is configured to perform fusion processing on each of the original map data by using an object matching the feature of the object as a reference object.
  • an embodiment of the present invention provides a non-volatile computer storage medium, where the computer storage medium stores one or more modules, when the one or more modules are processed by one high-precision map data.
  • the device of the method When executed, the device is caused to perform the following operations:
  • an embodiment of the present invention provides a device, where the device includes:
  • One or more processors are One or more processors;
  • One or more programs the one or more programs being stored in the memory, and when executed by the one or more processors, do the following:
  • Each of the original map data is subjected to fusion processing by using an object matching the feature of the object as a reference.
  • the method, device, storage medium and device for processing high-precision map data provided by the embodiments of the present invention, by identifying objects in the original map data, and extracting object features, and then matching and matching the object features in each original map data
  • the obtained object is used as a reference object, and the original map data is fused by the reference object, which solves the problem that the object in the map image is misaligned, removes the ghost problem of the map image, and improves the sharpness of the map image.
  • FIG. 1 is a flowchart of a method for processing high-precision map data according to Embodiment 1 of the present invention
  • Embodiment 2 is a schematic diagram of a three-digit coordinate system of a minimum outer package provided by Embodiment 2 of the present invention
  • FIG. 3 is a schematic structural diagram of a device for processing high-precision map data according to Embodiment 3 of the present invention.
  • FIG. 4 is a schematic structural diagram of hardware of a device for performing high-precision map data processing method according to Embodiment 5 of the present invention.
  • FIG. 1 is a flowchart of a method for processing high-precision map data according to Embodiment 1 of the present invention.
  • the method can be performed by a processing device of high-precision map data, wherein the device can be implemented by software and/or hardware, and can generally be integrated into a server having image data processing capabilities.
  • the method specifically includes the following:
  • the target trajectory may refer to a collection route of a collection vehicle for collecting map data, a collection nautical model, or a collection robot, and the route may be pre-set by an artificial person and stored in the collection tool. It can also be a better acquisition route that the collection tool determines itself based on the acquisition start point and end point of the human input.
  • the acquisition tool can collect the map data of the objects around the same acquisition track at least twice.
  • the advantage of this setting is that by performing fusion processing on the map data of the same object obtained in multiple acquisitions, the information of the object can be presented in the map as much as possible, thereby improving the user experience.
  • the collection tool includes at least one of an IMU, GPS or LIDAR acquisition mode.
  • the acquisition tool collects the map data of the surrounding objects along the target trajectory to obtain point cloud data and image data of the corresponding object.
  • the map data is processed by the collection tool.
  • the collection tool is used to identify the dynamic object or the static object in the original map data collected by each acquisition, for example, according to the original map data obtained by the acquisition.
  • the IMU can calculate the motion shape of a dynamic object
  • LIDAR can calculate the topography or structural features of a static object.
  • the features of the different objects are extracted.
  • features such as speed and/or acceleration of movement per unit time may be extracted; for a static object, structural features, center of gravity features, and/or topographic features may be extracted.
  • the object features of the static object are acquired for convenience as a reference feature.
  • the iconic object when identifying an object in each original map data, the iconic object may be selectively identified, or all objects may be identified.
  • the iconic object in each original map data is identified, for example, identifying a road plane in the original map data, a vehicle located on the road plane, a street light pole, a street sign or a building plane And other iconic objects.
  • the advantage of this arrangement is that on the one hand, the identification of the iconic object makes the recognition accuracy higher. On the other hand, only identifying the iconic object shortens the time required for the recognition and reduces the recognition cost.
  • the object in the original map data is identified by the collection tool, and the features of the objects identified in the original map data are extracted. After the extraction is completed, the phases of the original map data are Match the same object features.
  • the advantage of this setting is that the matching can eliminate the ghosting problem caused by the object features identified by each original map data in the fusion processing, thereby ensuring that the objects after matching the object features are consistent with the actual objects, thereby ensuring the accuracy of the map. Sex.
  • the Random Sample Consensus (RANSAC) algorithm may be further used to remove the mismatching operation in the matching process, so that the objects after matching the object features are The consistency of the actual objects is higher, which in turn ensures the accuracy of the map more effectively.
  • the object that matches the feature of the object is used as a reference object, and each of the original map data is subjected to fusion processing.
  • the object matching the feature of the object is used as a reference object, and the original map data representing the different objects are fused by the relative positional relationship of the reference object, and finally a high-precision map of the weight loss and deviation of the object is obtained.
  • the object features in each original map data are matched, and the matched object is used as a reference object.
  • the relative positional relationship of the reference object is used to fuse the original map data, which solves the problem of object misalignment occurring in the fusion processing of each original map data, thereby removing the ghosting problem of the map image and improving the map.
  • the sharpness of the image is used to fuse the original map data, which solves the problem of object misalignment occurring in the fusion processing of each original map data, thereby removing the ghosting problem of the map image and improving the map.
  • the method further comprises: separately analyzing and segmenting the original map data according to the collected trajectory; The segmented original map data paragraphs are associated as the associated original map data for subsequent recognition of the object.
  • the original map data obtained by each acquisition is analyzed, and the analyzed content may include the length of the collected trajectory, the gradation value of the collected original map data or the characteristic value of the iconic object, and the like;
  • the segmentation can be divided into several paragraphs according to the length of the collected trajectory or according to a certain rule, divided into several paragraphs according to the change of the gray value of the original map data or divided into several paragraphs according to the characteristic value of the iconic object.
  • the principle of the segmentation is not limited, and the segmentation principle may be performed according to the pre-set division principle or the default segmentation principle of the collection tool; in addition, the number of segments to be segmented is not limited.
  • the acquisition tool collects the offset rotation angle and the offset amount on the acquisition trajectory of different paragraphs during the process of collecting the map data of the objects around the acquisition trajectory, if the entire acquisition is directly
  • the value is set to a fixed value, only the offset rotation angle and the offset amount can be eliminated.
  • the advantage of analyzing and segmenting the collected trajectory is that after the original map data of the object is subsequently identified and the feature of the object is extracted, the original rotation angle and the offset of each original map data segment are used for each original.
  • the map data is fused to better solve the problem of misalignment of objects appearing in the fusion process, and remove the ghosting problem of the map image.
  • the object matching the feature of the object is used as a reference object, and the fusion processing of each of the original map data is optimized to calculate an object that matches the feature of the object as a reference object.
  • the offset rotation angle and the offset of the original map data; the original map data is subjected to fusion processing according to the offset rotation angle and the offset amount of each original map data.
  • the offset rotation angle and the offset of each original map data are calculated, wherein The offset rotation angle and the offset of the object matching the object feature are calculated, and the offset rotation angle and the offset of the object are taken as the offset rotation angle and the offset of the original map data segment corresponding thereto.
  • the original map data segments are merged according to the offset rotation angle and the offset of each original map data segment, and according to the determined reference object.
  • an offset angle and an offset of each original map data are calculated by an Iterative Closest Point (ICP) algorithm.
  • ICP Iterative Closest Point
  • the method for processing high-precision map data provided by the second embodiment of the present invention is based on the foregoing embodiment. Further, the object in the original map data is separately identified, and the feature of the object is extracted and optimized. The point cloud data in the original map data performs static object recognition, and extracts structural features of the static object.
  • Point cloud data includes reflectivity and topography of different objects. Since the point cloud data is a series of discrete points of the object data surface model, compared with the use of image data to identify the object, using the point cloud data to identify the static object can improve the accuracy of the recognized object, thereby improving the extracted The accuracy of the object features.
  • the lane line is identified by using the reflectivity in the point cloud data in the original map data, and the road plane is determined by lane line fitting, and the road is extracted. Plane planar features.
  • the acquisition tool identifies the lane line using the reflectivity in the point cloud data in the original map data. Since the lane line is made of special reflective paint, the reflectivity of the lane line is higher than that of other objects. Therefore, with its high reflectivity, the acquisition tool recognizes the lane line from each original map data. Further, After the lane line is identified, the road plane is determined by fitting the width and length of the lane line, and the plane features of the road plane are extracted, which may include features such as the width or length of the road plane.
  • a minimum outsourcing box identification is performed on an object that exceeds the road plane, and the vehicle and the street light pole are separately identified according to the characteristics of the minimum outer box, and the vehicle top plane feature of the stationary vehicle is extracted. And extracting the center of gravity feature of the street light pole.
  • the road plane can be determined in a number of ways, and then the street light pole is identified based on the road plane.
  • the outer casing may refer to an imaginary hexahedral structure capable of holding an object, and the smallest outer casing may refer to an imaginary hexahedral structure that can hold an object. According to the topography of the object, the characteristics of the smallest outer box of different objects are different.
  • the minimum outer package feature may refer to a three-dimensional coordinate A (x1, y1, z1) of a set apex A of the minimum outer package, wherein X1 represents the length of the smallest outer box, y1 represents the width of the smallest outer box, z1 represents the height of the smallest outer box, and may also mean the average of the distance from the center point 0 of the smallest outer box to each vertex.
  • the minimum outsourcing box feature is set as the three-dimensional coordinate A (x1, y1, z1) of the vertex A, and the minimum outer box of the vehicle
  • the characteristic is x1>z1, and y1>z1
  • the minimum outer box feature of the street light pole is z1>>x1, and z1>>y1.
  • the minimum outsourcing box of the object beyond the road plane is identified, and the iconic object vehicle and the street light pole are respectively identified according to the characteristics of the smallest outer box of different objects, and the identification vehicle is extracted.
  • the top feature of the car's top is extracted to identify the center of gravity of the lamppost.
  • the point cloud data with the reflectance reaching the high reflectivity condition is clustered, and the clustered point cloud data is plane-fitted and The street sign is identified in the fitted plane, and the planar features of the street sign are extracted.
  • LIDAR is used to collect the point data of the object to obtain the point cloud data of the object.
  • the point cloud data contains the reflectivity of different objects. Among the objects around the collected track, some objects have low reflectivity, such as vegetation, vehicles or buildings. Things, etc., other parts of the object have higher reflectivity, such as lane lines, glass or street signs.
  • the collection tool analyzes and calculates the reflectivity of different objects in the point cloud data. When the reflectance is higher than the preset threshold range, the point cloud data of the object is determined to be point cloud data of a high reflection condition, wherein the preset threshold range is It can be set by the person according to the corresponding acquisition trajectory in advance, or it can be set by the collection tool by default.
  • the point cloud data is classified, and the point cloud data of the same object whose reflectance reaches the high reflectivity condition is clustered, and the clustered point cloud data is plane-fitted.
  • the road sign is identified in the fitted plane, and the planar features of the identification street sign are extracted.
  • the street signs are located on both sides of the road plane or at intersections.
  • the plane features of the road signs are extracted, and the street sign plane features in each original map data are matched in subsequent operations, and the road signs matching the road sign plane features are used as reference
  • the accuracy of the fusion processing can be further improved, the ghosting problem of the map image can be better solved, and the sharpness of the map image can be improved.
  • the technical solution of the embodiment identifies the static object by using the point cloud data, and extracts the structural features of the static object. Further, according to the reflectivity in the point cloud data, the stationary vehicle and the street lamp are identified from the road plane. Map data of landmark objects such as poles and/or street signs, thus In the process, the iconic object features in each original map data are matched, and the iconic object matching the object feature is used as a reference object, and the original map data is merged by using the relative positional relationship of the iconic object. The accuracy of the fusion processing can be further improved, the ghosting problem of the map image can be better solved, and the sharpness of the map image can be improved.
  • FIG. 3 is a schematic structural diagram of a device for processing high-precision map data according to Embodiment 3 of the present invention.
  • the apparatus of this embodiment specifically includes: an acquisition module 31, an identification module 32, a matching module 33, and a fusion processing module 34;
  • the collecting module 31 is configured to perform at least two acquisitions on the map data along the target trajectory to obtain at least two original map data respectively.
  • the identification module 32 is configured to respectively identify an object in the original map data, and extract an object feature
  • a matching module 33 configured to match object features in each of the original map data
  • the fusion processing module 34 is configured to perform fusion processing on each of the original map data by using an object matching the feature of the object as a reference object.
  • the map data is collected at least twice along the target trajectory by the acquisition module 31 to acquire at least two original map data respectively, and the at least two objects in the original map data acquired by the identification module 32 are acquired.
  • the identification extracting the feature of the object, and then matching the feature of the object in each of the extracted original map data by the matching module 33.
  • the object after matching the feature of the object is used as a reference by the fusion processing module 34, and the reference is used.
  • the relative positional relationship of the objects is used to fuse the original map data.
  • the identification module 32 in the device is specifically configured to: perform static object recognition by using point cloud data in the original map data, and extract structural features of the static object.
  • the identification module 32 is specifically configured to: identify a lane line from the road plane according to a reflectivity in the point cloud data in the original map data, and fit the road plane through the lane line to extract the road plane Plane features.
  • the identification module 32 is specifically configured to: perform minimum outsourcing box identification on objects beyond the road plane according to the point cloud data, identify the vehicle and the street light pole according to the characteristics of the minimum outsourcing box, and extract the static A vehicle top plane feature of the vehicle and extracting a center of gravity feature of the street light pole.
  • the identification module 32 is specifically configured to: cluster point cloud data having a reflectance to a high reflectivity condition according to a reflectance in the point cloud data in the original map data, and cluster the point cloud after the clustering
  • the data is planarly fitted and the street signs are identified from the fitted planes to extract the planar features of the street signs.
  • the device further includes:
  • the analyzing and segmentation module 35 is configured to separately analyze and segment the original map data according to the collected trajectory before separately identifying the objects in the original map data and extracting the object features;
  • the association module 36 is configured to associate the segmented original map data segments as associated original map data of the subsequent identified objects.
  • the fusion processing module 34 includes:
  • the offset parameter calculation unit 341 is configured to calculate an offset rotation angle and an offset amount of each of the original map data by using an object matching the object feature as a reference object;
  • the fusion processing unit 342 is configured to perform fusion processing on the original map data according to the offset rotation angle and the offset of each original map data.
  • the processing device for the above-mentioned map data can execute the processing method of the high-precision map data provided by any embodiment of the present invention, and has the functional modules and the beneficial effects corresponding to the execution method.
  • the embodiment provides a non-volatile computer storage medium storing one or more modules when the one or more modules are executed by a device that executes a processing method of high-precision map data. , causing the device to perform the following operations:
  • Each of the original map data is subjected to fusion processing by using an object matching the feature of the object as a reference.
  • the separately identifying the object in the original map data and extracting the feature of the object may include:
  • Static object recognition is performed by using point cloud data in the original map data, and structural features of the static object are extracted.
  • the performing the static object recognition by using the point cloud data in the original map data, and extracting the structural features of the static object may preferably include:
  • a lane line is identified according to a reflectance in the point cloud data in the original map data, and a plane feature of the road plane is extracted by fitting the road plane through the lane line.
  • the static object recognition is performed by using the point cloud data in the original map data, and the structural features of the static object are extracted, and the method further includes:
  • the static object recognition is performed by using the point cloud data in the original map data, and the structural features of the static object are extracted, and the method further includes:
  • the point cloud data with the reflectance reaching the high reflectivity condition is clustered, and the clustered point cloud data is plane-fitted and fitted from the fitting
  • the street sign is identified in the plane, and the planar features of the street sign are extracted.
  • the device may further include:
  • the original map data is separately analyzed and segmented according to the collected trajectory
  • the segmented original map data paragraphs are associated as the associated original map data of the subsequent identified objects.
  • the object that matches the feature of the object is used as a reference object, and the original map data is subjected to fusion processing, which may preferably include:
  • the original map data is subjected to fusion processing according to the offset rotation angle and the offset amount of each original map data.
  • FIG. 4 is a schematic structural diagram of hardware of a device for performing high-precision map data processing method according to Embodiment 5 of the present invention.
  • the device includes:
  • One or more processors 410, one processor 410 is taken as an example in FIG. 4;
  • Memory 420 and one or more modules.
  • the device may also include an input device 430 and an output device 440.
  • the processor 410, the memory 420, the input device 430, and the output device 440 in the device may be connected by a bus or other means, and the bus connection is taken as an example in FIG.
  • the memory 420 is used as a computer readable storage medium, and can be used to store software programs, computer executable programs, and modules, such as program instructions/modules corresponding to the processing method of high-precision map data in the embodiment of the present invention (for example, FIG. 3
  • the processor 410 executes various functional applications and data processing of the server by executing software programs, instructions, and modules stored in the memory 420, that is, a method for processing high-precision map data of the above method embodiments.
  • the memory 420 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application required for at least one function; the storage data area may store data created according to usage of the terminal device, and the like.
  • memory 420 can include high speed random access memory, and can also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device.
  • memory 420 can further include memory remotely located relative to processor 410, which can be connected to the terminal device over a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.
  • Input device 430 can be used to receive input digital or character information and to generate key signal inputs related to user settings and function control of the terminal.
  • Output device 440 can include a display device such as a display screen.
  • the one or more modules are stored in the memory 420, and when executed by the one or more processors 410, perform the following operations:
  • Each of the original map data is subjected to fusion processing by using an object matching the feature of the object as a reference.
  • the separately identifying the objects in the original map data and extracting the object features may include:
  • Static object recognition is performed by using point cloud data in the original map data, and structural features of the static object are extracted.
  • the performing the static object recognition by using the point cloud data in the original map data, and extracting the structural features of the static object may further include:
  • a lane line is identified according to a reflectance in the point cloud data in the original map data, and a plane feature of the road plane is extracted by fitting the road plane through the lane line.
  • the performing the static object recognition by using the point cloud data in the original map data, and extracting the structural features of the static object may further include:
  • the performing the static object recognition by using the point cloud data in the original map data, and extracting the structural features of the static object may further include:
  • the point cloud data with the reflectance reaching the high reflectivity condition is clustered, and the clustered point cloud data is plane-fitted and fitted from the fitting
  • the street sign is identified in the plane, and the planar features of the street sign are extracted.
  • the method may further include:
  • the original map data is separately analyzed and segmented according to the collected trajectory
  • the segmented original map data paragraphs are associated as the associated original map data of the subsequent identified objects.
  • the object that matches the feature of the object is used as a reference object, and the original map data is subjected to fusion processing, which may include:
  • the original map data is subjected to fusion processing according to the offset rotation angle and the offset amount of each original map data.
  • the present invention can be implemented by software and necessary general hardware, and can also be implemented by hardware, but in many cases, the former is a better implementation. .
  • the technical solution of the present invention which is essential or contributes to the prior art, may be embodied in the form of a software product, which may be stored in a computer readable storage medium, such as a floppy disk of a computer. , Read-Only Memory (ROM), Random Access Memory (RAM), Flash (FLASH), hard disk or optical disk, etc., including a number of instructions to make a computer device (can be a personal computer)
  • the server, or network device, etc. performs the methods described in various embodiments of the present invention.
  • the units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be implemented.
  • the specific names of the respective functional units are only for convenience of distinguishing from each other, and are not intended to limit the scope of protection of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Remote Sensing (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Geometry (AREA)
  • Software Systems (AREA)
  • Evolutionary Biology (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Graphics (AREA)
  • Educational Administration (AREA)
  • Educational Technology (AREA)
  • Business, Economics & Management (AREA)
  • Mathematical Physics (AREA)
  • Navigation (AREA)
  • Image Analysis (AREA)
  • Instructional Devices (AREA)
  • Processing Or Creating Images (AREA)

Abstract

一种高精地图数据的处理方法、装置、存储介质和设备,该方法包括:沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据(S101);分别识别所述原始地图数据中的物体,并提取物体特征(S102);将各所述原始地图数据中的物体特征进行匹配(S103);将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理(S104)。该方法通过识别原始地图数据中的物体,并提取物体特征,进而将各原始地图数据中的物体特征进行匹配并将匹配后得到的物体作为参照物,利用该参照物对原始地图数据进行融合处理,解决了地图图像中的物体出现错位的问题,去除了地图图像的重影问题,提高了地图图像的清晰度。

Description

高精地图数据的处理方法、装置、存储介质和设备
本专利申请要求于2015年9月9日提交的、申请号为201510571925.6、申请人为百度在线网络技术(北京)有限公司、发明名称为“一种高精地图数据的处理方法和装置”的中国专利申请的优先权,该申请的全文以引用的方式并入本申请中。
技术领域
本发明实施例涉及地图数据处理技术,尤其涉及高精地图数据的处理方法、装置、存储介质和设备。
背景技术
随着自驾出行探亲、探险和旅游的用户越来越多,当用户去往自己不熟悉的目的地时,便会依赖于利用智能终端进行地图查询,用户可以在智能终端中通过输入目的地对出行路线进行查看和选择,地图的使用在很大程度上方便了人们的生活。
目前,形成地图图像之前,对地图图像数据的采集一般利用惯性测量单元(Intertial Measurement Uint,IMU)、全球定位系统(Global Positioning System,GPS)或者光探测与测量(Light Detection And Ranging,LIDAR)等多种工具来完成,再将数据进行融合处理,形成地图数据。
IMU主要包括陀螺仪和加速度计两部分,该两部分直接影响IMU的测量图像数据的精度。其中陀螺仪主要测量运动物体的角速度信息,加速度计主要测量运动物体的加速度信息,IMU根据所测信息计算得到运动物体的运动形态。
GPS主要包括空间星座、地面监控和用户设备三部分,该三部分直接影响GPS获得的图像数据的精度。其中,空间星座一般由24颗卫星组成,用于实时测量地面物体的图像数据,地面监控接收空间星座的卫星发送的该卫星测量得到的物体的地图数据,用户设备对该物体的地图数据进行计算并得到物体的位置信息等。
LIDAR主要利用IMU和GPS进行激光扫描,其所测得的数据为点云数据,即 所测得的数据为物体表面模型的一系列离散点,该点云数据中包括物体的空间三维信息和激光强度信息。LIDAR中的传感器将激光束发射到地面或者物体表面,激光束遇到障碍物发生发射,激光束反射的能量被该传感器记录并作为激光强度信息,并且利用该激光强度信息可以计算得到激光束的反射率。
通常为了保证测量得到的地图数据的多样性和全面性,利用上述工具对获得地图数据的采集轨迹进行多次采集。地图数据采集结束后,利用GPS或IMU对采集得到的多次相应地图数据进行关联,最终得到高精地图。
然而,现有技术存在以下缺陷,在利用GPS或IMU对采集得到的多次相应的地图数据进行关联时,由于工具本身的测量和计算精度的限制,使得关联后得到的地图图像中的物体出现错位,导致用户看到的地图图像中的物体有重影,严重影响用户的使用体验。
发明内容
本发明实施例的主要目的在于,提供高精地图数据的处理方法、装置、存储介质和设备。以实现解决对多次采集的图像数据进行关联时出现的地图图像中的物体错位和重影的问题。
本发明实施例采用如下技术方案:
第一方面,本发明实施例提供了一种高精地图数据的处理方法,该方法包括:
沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
分别识别所述原始地图数据中的物体,并提取物体特征;
将各所述原始地图数据中的物体特征进行匹配;
将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
第二方面,本发明实施例还提供了一种高精地图数据的处理装置,该装置包括:
采集模块,用于沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
识别模块,用于分别识别所述原始地图数据中的物体,并提取物体特征;
匹配模块,用于将各所述原始地图数据中的物体特征进行匹配;
融合处理模块,用于将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
第三方面,本发明实施例提供了一种非易失性计算机存储介质,所述计算机存储介质存储有一个或者多个模块,当所述一个或者多个模块被一个执行高精地图数据的处理方法的设备执行时,使得所述设备执行如下操作:
沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
分别识别所述原始地图数据中的物体,并提取物体特征;
将各所述原始地图数据中的物体特征进行匹配;
将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。第四方面,本发明实施例提供了一种设备,该设备包括:
一个或多个处理器;
存储器;
一个或多个程序,所述一个或多个程序存储在所述存储器中,当被所述一个或多个处理器执行时,进行如下操作:
沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
分别识别所述原始地图数据中的物体,并提取物体特征;
将各所述原始地图数据中的物体特征进行匹配;
将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
本发明实施例提供的高精地图数据的处理方法、装置、存储介质和设备,通过识别原始地图数据中的物体,并提取物体特征,进而将各原始地图数据中的物体特征进行匹配并将匹配后得到的物体作为参照物,利用该参照物对原始地图数据进行融合处理,解决了地图图像中的物体出现错位的问题,去除了地图图像的重影问题,提高了地图图像的清晰度。
附图说明
为了更清楚地说明本发明实施例中的技术方案,下面将对实施例描述中所需使用的附图作简单地介绍,当然,以下描述中的附图仅仅是本发明的一些实施例,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,还可以对这些附图进行修改和替换。
图1为本发明实施例一提供的一种高精地图数据的处理方法的流程图;
图2为本发明实施例二提供的最小外包盒的三位坐标系示意图;
图3为本发明实施例三提供的一种高精地图数据的处理装置的结构示意图。
图4为本发明实施例五提供的一种执行高精地图数据的处理方法的设备的硬件结构示意图。
具体实施方式
下面将结合附图对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例是本发明的一部分实施例,而不是全部的实施例,是为了阐述本发明的原理,而不是要将本发明限制于这些具体的实施例。基于本发明中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本发明保护的范围。
实施例一
图1为本发明实施例一提供的一种高精地图数据的处理方法的流程图。该方法可以由高精地图数据的处理装置执行,其中该装置可由软件和/或硬件实现,并一般可集成在具有图像数据处理能力的服务器中。
参见图1,该方法具体包括如下:
S101、沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据。
在上述操作中,所述目标轨迹可以指的是进行地图数据采集的采集车、采集航模或者采集机器人等采集工具的采集路线,该路线可以是人为预先设定并存储在上述采集工具中的,也可以是采集工具根据人为输入的采集起始点和终止点自行确定的较佳的采集路线。
通常,采集工具在对目标轨迹周围物体的地图数据进行采集时,不能保证一次完成所有物体的地图数据的完整采集,为了保证采集得到的目标轨迹周围物体的地图数据的多样性和全面性,一般情况下,采集工具可以对同一采集轨迹周围物体的地图数据进行至少两次采集。这样设置的好处是,通过对多次采集中得到的同一物体的地图数据进行融合处理,可以使得该物体尽可能多的信息呈现在地图中,提升用户的使用体验。
在此,需要进一步说明的是,采集工具中至少包括IMU、GPS或LIDAR采集方式中的一种。采集工具沿目标轨迹对周围物体的地图数据进行采集,得到相应物体的点云数据和图像数据。
S102、分别识别所述原始地图数据中的物体,并提取物体特征。
上述操作中,在地图数据采集完毕之后,利用采集工具对地图数据进行处理,首先利用采集工具识别各次采集得到的原始地图数据中的动态物体或静态物体,例如根据采集得到的原始地图数据,IMU可以计算出动态物体的运动形态,LIDAR可以计算出静态物体的形貌特征或者结构特征。识别到原始地图数据中的物体后,提取不同物体的特征。优选的,对于动态物体而言,可以提取其单位时间内移动的速度和/或加速度等特征;对于静态物体而言,可以提取其结构特征、重心特征和/或形貌特征等。优选是采集静态物体的物体特征,以便于作为参照特征。
在此,需要说明的是,对各次原始地图数据中的物体进行识别时,可以选择性地对具有标志性的物体进行识别,也可以对全部物体均进行识别。在本实施例中,优选的,对各次原始地图数据中具有标志性的物体进行识别,例如,识别原始地图数据中的道路平面、位于道路平面上的车辆、路灯杆、路牌或者建筑物平面等标志性物体。这样设置的好处是,一方面对具有标志性的物体进行识别使得识别的精准度提高,另一方面,只识别具有标志性的物体缩短了识别所需要的时间,降低了识别成本。
S103、将各所述原始地图数据中的物体特征进行匹配。
上述操作中,利用采集工具识别各次原始地图数据中物体,并提取各次原始地图数据中识别的物体的特征。提取完毕后,将所述各次原始地图数据中相 同的物体特征进行匹配。这样设置的好处是,经过匹配可以消除各次原始地图数据识别得到的物体特征在融合处理中产生的重影问题,进而保证匹配物体特征之后的物体与实际物体保持一致,进而保证了地图的精准性。
优选的,将各次原始地图数据中的物体特征进行匹配之后,还可以进一步采用随机抽样一致性(Random Sample Consensus,RANSAC)算法去除匹配过程中的误匹配操作,使得匹配物体特征后的物体与实际物体的一致性更高,进而更有效地保证了地图的精准性。
S104、将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
上述操作中,将匹配物体特征后的物体作为参照物,利用参照物的相对位置关系,将代表不同物体的各次原始地图数据进行融合处理,最终得到消除物体重影和偏差的高精地图。
本实施例提供的技术方案,通过识别原始地图数据中的物体,并提取识别物体的特征,进而将各次原始地图数据中的物体特征进行匹配,并将匹配后得到的物体作为参照物,利用该参照物的相对位置关系对各次原始地图数据进行融合处理,解决了在对各次原始地图数据进行融合处理中出现的物体错位的问题,从而去除了地图图像的重影问题,提高了地图图像的清晰度。
在上述技术方案的基础上,优选的,在所述分别识别所述原始地图数据中的物体,并提取物体特征之前,还包括分别对所述原始地图数据按照采集轨迹进行分析和切分;将切分后的原始地图数据段落进行关联,作为后续识别物体的关联原始地图数据。
首先对各次采集得到的原始地图数据进行分析,分析的内容可以包括采集轨迹的长度、采集得到的原始地图数据的灰度值或标志性物体的特征值等;然后对各次原始地图数据进行切分,切分可以按照采集轨迹的长度平均或按一定规律切分成若干段落、按照原始地图数据的灰度值的变化切分成若干段落或者按照标志性物体的特征值切分成若干段落。在此,对切分的原则不作任何限定,可以按照人为预先设定的切分原则或者采集工具默认的切分原则进行切分;另外,对切分的段落数量也不作任何限定。
将切分后的相应的各次原始地图数据段落进行关联,并分析关联之后各次原始地图数据段落的偏移旋转角度和偏移量,以方便后续的操作。
当采集轨迹较长时,采集工具在对采集轨迹周围的物体进行地图数据采集的过程中,在不同段落的采集轨迹上采集的偏移旋转角度和偏移量有所不同,若直接对整个采集轨迹进行处理,那么在对各次原始地图数据进行融合处理时,难以统一设定纠偏旋转角度和纠偏量,若设定为固定值,那么只能消除与该偏移旋转角度和偏移量相同或接近的原始地图数据段落中地图数据的偏差,解决其地图图像的重影问题,而其他与该偏移旋转角度和偏移量相差较大的原始地图数据段落中的地图数据的偏差仍然存在,地图图像的重影问题也将存在。
因此,将采集轨迹进行分析和切分的好处在于,在后续识别物体的关联原始地图数据,并提取物体特征之后,按照各次原始地图数据段落的偏移旋转角度和偏移量对各次原始地图数据进行融合处理,更好地解决融合处理中出现的物体错位的问题,去除地图图像的重影问题。
进一步的,在上述技术方案的基础上,所述将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理优化为,将匹配物体特征的物体作为参照物,计算各所述原始地图数据的偏移旋转角度和偏移量;根据各原始地图数据的偏移旋转角度和偏移量,对原始地图数据进行融合处理。
对各次原始地图数据按照采集轨迹进行分析和切分,并将切分后的各次原始地图数据段落进行关联之后,计算各次原始地图数据的偏移旋转角度和偏移量,其中,可以计算匹配过物体特征的物体的偏移旋转角度和偏移量,并将该物体的偏移旋转角度和偏移量作为与之相对应的原始地图数据段落的偏移旋转角度和偏移量。
根据各次原始地图数据段落的偏移旋转角度和偏移量,以及根据确定的参照物,对各次原始地图数据段落进行融合处理。
优选的,采用迭代就近点(Iterative Closest Point,ICP)算法计算各次原始地图数据的偏移量角度和偏移量。
实施例二
本发明实施例二提供的一种高精地图数据的处理方法以上述实施例为基础,进一步的,对所述分别识别所述原始地图数据中的物体,并提取物体特征进行优化,优化为采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征。
点云数据中包括不同物体的反射率和形貌特征。由于点云数据为物体数据表面模型的一系列离散点,因此与利用图像数据对物体进行识别相比,利用点云数据对静态物体进行识别,可以提高识别物体的精确度,进而提高提取得到的物体特征的精确度。
可选的,在上述各实施例的技术方案的基础上,利用所述原始地图数据中的点云数据中的反射率,识别车道线,并通过车道线拟合确定道路平面,提取所述道路平面的平面特征。
采集工具利用原始地图数据中的点云数据中的反射率识别车道线。由于车道线采用特殊的反光漆制作,使得车道线的反射率高于其他物体的反射率,因此,利用其高的反射率,采集工具从各次原始地图数据中识别出车道线,进一步的,在识别出车道线后,通过车道线的宽度和长度拟合确定道路平面,并提取该道路平面的平面特征,该平面特征可以包括道路平面的宽度或长度等特征。
进一步的,根据所述点云数据,对超出所述道路平面的物体进行最小外包盒识别,根据所述最小外包盒的特征分别识别车辆和路灯杆,提取所述静止车辆的车顶部平面特征,且提取所述路灯杆的重心特征。
道路平面可以多种方式进行确定,随后再基于道路平面识别路灯杆。所述外包盒可以是指能盛装物体的假想六面体结构,最小外包盒可以是指恰好能盛装物体的假想六面体结构。根据物体的形貌特征,不同物体的最小外包盒特征不同。
图2为本发明实施例二提供的最小外包盒的示意图,如图2所示,最小外包盒特征可以是指最小外包盒的设定顶点A的三维坐标A(x1,y1,z1),其中,x1表示最小外包盒的长度,y1表示最小外包盒的宽度,z1表示最小外包盒的高度;也可以是指最小外包盒的中心点0到各个顶点的距离的平均值。示例性的,以最小外包盒特征为设定顶点A的三维坐标A(x1,y1,z1),车辆的最小外包盒 特征为x1>z1,且y1>z1,路灯杆的最小外包盒特征为z1>>x1,且z1>>y1。
根据采集工具采集得到的原始地图数据中的点云数据,识别超出道路平面的物体的最小外包盒,根据不同物体的最小外包盒的特征,分别识别标志性物体车辆和路灯杆,并提取识别车辆的车顶部平面特征,提取识别路灯杆的重心特征。
进一步的,根据所述原始地图数据中的点云数据中的反射率,将反射率达到高反射率条件的点云数据进行聚类,将聚类后的点云数据进行平面拟合,并从拟合的平面中识别路牌,提取所述路牌的平面特征。
利用LIDAR对物体的地图数据进行采集得到物体的点云数据,其中点云数据中包含不同物体的反射率,位于采集轨迹周围的物体中,一部分物体的反射率较低,例如草木、车辆或建筑物等,另一部分物体的反射率较高,例如车道线、玻璃或路牌等。采集工具对点云数据中的不同物体的反射率进行分析计算,当反射率高于预设阈值范围时,则确定该物体的点云数据为高反射条件的点云数据,其中预设阈值范围可以是人为预先根据相应采集轨迹进行设定,也可以是采集工具默认设定的。
根据反射率的高低对点云数据进行分类,并将上述确定的反射率达到高反射率条件的同一物体的点云数据进行聚类,并将聚类后的点云数据进行平面拟合,从拟合得到的平面中识别路牌,并提取识别路牌的平面特征。通常路牌位于道路平面的两侧或者十字路口,因此,提取路牌的平面特征,在后续操作中将各次原始地图数据中的路牌平面特征进行匹配,并将该匹配过路牌平面特征的路牌作为参照物,利用该路牌的相对位置关系对各个原始地图数据进行融合处理,可以进一步提高融合处理的精确度,更好地解决地图图像的重影问题,提高地图图像的清晰度。
在此,需要说明的是,上述根据点云数据中的反射率识别标志性物体的处理过程,可以单独采用,也可以结合采用。
本实施例的技术方案,通过采用点云数据对静态物体进行识别,并提取静态物体的结构特征,进一步的,根据点云数据中的反射率的高低,从道路平面中识别出静止车辆、路灯杆和/或路牌等标志性物体的地图数据,从而在后续操 作中将各次原始地图数据中的标志性物体特征进行匹配,并将该匹配过物体特征的标志性物体作为参照物,利用该标志性物体的相对位置关系对各个原始地图数据进行融合处理,可以进一步提高融合处理的精确度,更好地解决地图图像的重影问题,提高地图图像的清晰度。
实施例三
图3为本发明实施例三提供的一种高精地图数据的处理装置的结构示意图。
参见图3,该实施例的装置具体包括:采集模块31、识别模块32、匹配模块33和融合处理模块34;
其中,采集模块31,用于沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
识别模块32,用于分别识别所述原始地图数据中的物体,并提取物体特征;
匹配模块33,用于将各所述原始地图数据中的物体特征进行匹配;
融合处理模块34,用于将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
本实施例的技术方案,通过采集模块31沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据,并通过识别模块32对获取的至少两次原始地图数据中的物体进行识别,提取物体特征,然后通过匹配模块33将提取得到的各次原始地图数据中的物体特征进行匹配,匹配完成后,通过融合处理模块34将匹配物体特征后的物体作为参照物,利用参照物的相对位置关系对各次原始地图数据进行融合处理。从而解决了在对各次原始地图数据进行融合处理中出现的物体错位的问题,从而去除了地图图像的重影问题,提高了地图图像的清晰度。
进一步的,所述装置中的识别模块32具体用于:采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征。
进一步的,识别模块32具体用于:根据所述原始地图数据中的点云数据中的反射率,从道路平面中识别车道线,并通过所述车道线拟合道路平面,提取所述道路平面的平面特征。
进一步的,识别模块32具体用于:根据所述点云数据,对超出所述道路平面的物体进行最小外包盒识别,根据所述最小外包盒的特征分别识别车辆和路灯杆,提取所述静止车辆的车顶部平面特征,且提取所述路灯杆的重心特征。
更进一步的,识别模块32具体用于:根据所述原始地图数据中的点云数据中的反射率,将反射率达到高反射率条件的点云数据进行聚类,将聚类后的点云数据进行平面拟合,并从拟合的平面中识别路牌,提取所述路牌的平面特征。
上述技术方案基础上,该装置还包括:
分析和切分模块35,用于在分别识别所述原始地图数据中的物体,并提取物体特征之前,分别对所述原始地图数据按照采集轨迹进行分析和切分;
关联模块36,用于将切分后的原始地图数据段落进行关联,作为后续识别物体的关联原始地图数据。
进一步的,所述融合处理模块34包括:
偏移参数计算单元341,用于将匹配物体特征的物体作为参照物,计算各所述原始地图数据的偏移旋转角度和偏移量;
融合处理单元342,用于根据各原始地图数据的偏移旋转角度和偏移量,对原始地图数据进行融合处理。
上述地图数据的处理装置可执行本发明任意实施例所提供的高精地图数据的处理方法,具备执行方法相应的功能模块和有益效果。
实施例四
本实施例提供了一种非易失性计算机存储介质,所述计算机存储介质存储有一个或者多个模块,当所述一个或者多个模块被一个执行高精地图数据的处理方法的设备执行时,使得所述设备执行如下操作:
沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
分别识别所述原始地图数据中的物体,并提取物体特征;
将各所述原始地图数据中的物体特征进行匹配;
将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
上述存储介质中存储的模块被所述设备所执行时,所述分别识别所述原始地图数据中的物体,并提取物体特征,可优选包括:
采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征。
上述存储介质中存储的模块被所述设备所执行时,所述采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征,可优选包括:
根据所述原始地图数据中的点云数据中的反射率,识别车道线,并通过所述车道线拟合道路平面,提取所述道路平面的平面特征。
上述存储介质中存储的模块被所述设备所执行时,所述采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征,还可优选包括:
根据所述点云数据,对超出道路平面的物体进行最小外包盒识别,根据所述最小外包盒的特征分别识别车辆和路灯杆,提取所述静止车辆的车顶部平面特征,且提取所述路灯杆的重心特征。
上述存储介质中存储的模块被所述设备所执行时,所述采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征,还可优选包括:
根据所述原始地图数据中的点云数据中的反射率,将反射率达到高反射率条件的点云数据进行聚类,将聚类后的点云数据进行平面拟合,并从拟合的平面中识别路牌,提取所述路牌的平面特征。
上述存储介质中存储的模块被所述设备所执行时,所述分别识别所述原始地图数据中的物体,并提取物体特征之前,还可包括:
分别对所述原始地图数据按照采集轨迹进行分析和切分;
将切分后的原始地图数据段落进行关联,作为后续识别物体的关联原始地图数据。
上述存储介质中存储的模块被所述设备所执行时,所述将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理,可优选包括:
将匹配物体特征的物体作为参照物,计算各所述原始地图数据的偏移旋转角度和偏移量;
根据各原始地图数据的偏移旋转角度和偏移量,对原始地图数据进行融合处理。
实施例五
图4为本发明实施例五提供的一种执行高精地图数据的处理方法的设备的硬件结构示意图。
参见图4,该设备包括:
一个或多个处理器410,图4中以一个处理器410为例;
存储器420;以及一个或者多个模块。
所述设备还可以包括:输入装置430和输出装置440。所述设备中的处理器410、存储器420、输入装置430和输出装置440可以通过总线或者其他方式连接,图4中以通过总线连接为例。
存储器420作为一种计算机可读存储介质,可用于存储软件程序、计算机可执行程序以及模块,如本发明实施例中的高精地图数据的处理方法对应的程序指令/模块(例如,附图3所述的采集模块31、识别模块32、匹配模块33和融合处理模块34)。处理器410通过运行存储在存储器420中的软件程序、指令以及模块,从而执行服务器的各种功能应用以及数据处理,即实现上述方法实施例的高精地图数据的处理方法。
存储器420可以包括存储程序区和存储数据区,其中,存储程序区可存储操作系统、至少一个功能所需要的应用程序;存储数据区可存储根据终端设备的使用所创建的数据等。此外,存储器420可以包括高速随机存取存储器,还可以包括非易失性存储器,例如至少一个磁盘存储器件、闪存器件、或其他非易失性固态存储器件。在一些实施例中,存储器420可进一步包括相对于处理器410远程设置的存储器,这些远程存储器可以通过网络连接至终端设备。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
输入装置430可用于接收输入的数字或字符信息,以及产生与终端的用户设置以及功能控制有关的键信号输入。输出装置440可包括显示屏等显示设备。
所述一个或者多个模块存储在所述存储器420中,当被所述一个或者多个处理器410执行时,执行如下操作:
沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
分别识别所述原始地图数据中的物体,并提取物体特征;
将各所述原始地图数据中的物体特征进行匹配;
将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
进一步的,所述分别识别所述原始地图数据中的物体,并提取物体特征,可包括:
采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征。
更进一步的,所述采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征,还可包括:
根据所述原始地图数据中的点云数据中的反射率,识别车道线,并通过所述车道线拟合道路平面,提取所述道路平面的平面特征。
更进一步的,所述采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征,还可包括:
根据所述点云数据,对超出道路平面的物体进行最小外包盒识别,根据所述最小外包盒的特征分别识别车辆和路灯杆,提取所述静止车辆的车顶部平面特征,且提取所述路灯杆的重心特征。
更进一步的,所述采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征,还可包括:
根据所述原始地图数据中的点云数据中的反射率,将反射率达到高反射率条件的点云数据进行聚类,将聚类后的点云数据进行平面拟合,并从拟合的平面中识别路牌,提取所述路牌的平面特征。
进一步的,所述分别识别所述原始地图数据中的物体,并提取物体特征之 前,所述方法还可包括:
分别对所述原始地图数据按照采集轨迹进行分析和切分;
将切分后的原始地图数据段落进行关联,作为后续识别物体的关联原始地图数据。
进一步的,所述将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理,可包括:
将匹配物体特征的物体作为参照物,计算各所述原始地图数据的偏移旋转角度和偏移量;
根据各原始地图数据的偏移旋转角度和偏移量,对原始地图数据进行融合处理。
通过以上关于实施方式的描述,所属领域的技术人员可以清楚地了解到,本发明可借助软件及必需的通用硬件来实现,当然也可以通过硬件实现,但很多情况下前者是更佳的实施方式。基于这样的理解,本发明的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品可以存储在计算机可读存储介质中,如计算机的软盘、只读存储器(Read-Only Memory,ROM)、随机存取存储器(Random Access Memory,RAM)、闪存(FLASH)、硬盘或光盘等,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行本发明各个实施例所述的方法。
值得注意的是,上述高精地图数据的处理装置的实施例中,所包括的各个单元和模块只是按照功能逻辑进行划分的,但并不局限于上述的划分,只要能够实现相应的功能即可;另外,各功能单元的具体名称也只是为了便于相互区分,并不用于限制本发明的保护范围。
以上所述,仅为本发明的具体实施方式,但本发明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发明披露的技术范围内,可轻易想到的变化或替换,都应涵盖在本发明的保护范围之内。因此,本发明的保护范围应以所述权利要求的保护范围为准。

Claims (16)

  1. 一种高精地图数据的处理方法,其特征在于,包括:
    沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
    分别识别所述原始地图数据中的物体,并提取物体特征;
    将各所述原始地图数据中的物体特征进行匹配;
    将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
  2. 根据权利要求1所述的方法,其特征在于,分别识别所述原始地图数据中的物体,并提取物体特征包括:
    采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征。
  3. 根据权利要求2所述的方法,其特征在于,采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征包括:
    根据所述原始地图数据中的点云数据中的反射率,识别车道线,并通过所述车道线拟合道路平面,提取所述道路平面的平面特征。
  4. 根据权利要求2所述的方法,其特征在于,采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征包括:
    根据所述点云数据,对超出道路平面的物体进行最小外包盒识别,根据所述最小外包盒的特征分别识别车辆和路灯杆,提取所述静止车辆的车顶部平面特征,且提取所述路灯杆的重心特征。
  5. 根据权利要求2所述的方法,其特征在于,采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征包括:
    根据所述原始地图数据中的点云数据中的反射率,将反射率达到高反射率条件的点云数据进行聚类,将聚类后的点云数据进行平面拟合,并从拟合的平面中识别路牌,提取所述路牌的平面特征。
  6. 根据权利要求1-5任一所述的方法,其特征在于,分别识别所述原始地图数据中的物体,并提取物体特征之前,还包括:
    分别对所述原始地图数据按照采集轨迹进行分析和切分;
    将切分后的原始地图数据段落进行关联,作为后续识别物体的关联原始地 图数据。
  7. 根据权利要求1-5任一所述的方法,其特征在于,将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理包括:
    将匹配物体特征的物体作为参照物,计算各所述原始地图数据的偏移旋转角度和偏移量;
    根据各原始地图数据的偏移旋转角度和偏移量,对原始地图数据进行融合处理。
  8. 一种高精地图数据的处理装置,其特征在于,包括:
    采集模块,用于沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
    识别模块,用于分别识别所述原始地图数据中的物体,并提取物体特征;
    匹配模块,用于将各所述原始地图数据中的物体特征进行匹配;
    融合处理模块,用于将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
  9. 根据权利要求8所述的装置,其特征在于,所述识别模块具体用于:
    采用所述原始地图数据中的点云数据进行静态物体识别,并提取所述静态物体的结构特征。
  10. 根据权利要求9所述的装置,其特征在于,所述识别模块具体用于:
    根据所述原始地图数据中的点云数据中的反射率,识别车道线,并通过所述车道线拟合道路平面,提取所述道路平面的平面特征。
  11. 根据权利要求9所述的装置,其特征在于,所述识别模块具体用于:
    根据所述点云数据,对超出道路平面的物体进行最小外包盒识别,根据所述最小外包盒的特征分别识别车辆和路灯杆,提取所述静止车辆的车顶部平面特征,且提取所述路灯杆的重心特征。
  12. 根据权利要求9所述的装置,其特征在于,所述识别模块具体用于:
    根据所述原始地图数据中的点云数据中的反射率,将反射率达到高反射率条件的点云数据进行聚类,将聚类后的点云数据进行平面拟合,并从拟合的平面中识别路牌,提取所述路牌的平面特征。
  13. 根据权利要求8-12任一所述的装置,其特征在于,还包括:
    分析和切分模块,用于在分别识别所述原始地图数据中的物体,并提取物体特征之前,分别对所述原始地图数据按照采集轨迹进行分析和切分;
    关联模块,用于将切分后的原始地图数据段落进行关联,作为后续识别物体的关联原始地图数据。
  14. 根据权利要求8-12任一所述的装置,其特征在于,所述融合处理模块包括:
    偏移参数计算单元,用于将匹配物体特征的物体作为参照物,计算各所述原始地图数据的偏移旋转角度和偏移量;
    融合处理单元,用于根据各原始地图数据的偏移旋转角度和偏移量,对原始地图数据进行融合处理。
  15. 一种非易失性计算机存储介质,所述计算机存储介质存储有一个或者多个模块,其特征在于,当所述一个或者多个模块被一个执行高精地图数据的处理方法的设备执行时,使得所述设备执行如下操作:
    沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
    分别识别所述原始地图数据中的物体,并提取物体特征;
    将各所述原始地图数据中的物体特征进行匹配;
    将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
  16. 一种设备,其特征在于,包括:
    一个或多个处理器;
    存储器;
    一个或多个程序,所述一个或多个程序存储在所述存储器中,当被所述一个或多个处理器执行时,进行如下操作:
    沿目标轨迹对地图数据进行至少两次采集,以分别获取至少两次原始地图数据;
    分别识别所述原始地图数据中的物体,并提取物体特征;
    将各所述原始地图数据中的物体特征进行匹配;
    将匹配物体特征的物体作为参照物,将各所述原始地图数据进行融合处理。
PCT/CN2015/099012 2015-09-09 2015-12-25 高精地图数据的处理方法、装置、存储介质和设备 Ceased WO2017041390A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
KR1020187007384A KR102037820B1 (ko) 2015-09-09 2015-12-25 고정밀 지도 데이터 처리 방법, 장치, 저장 매체 및 기기
JP2018512859A JP6664470B2 (ja) 2015-09-09 2015-12-25 高精度地図データの処理方法、装置、記憶媒体及び機器
US15/758,623 US10628712B2 (en) 2015-09-09 2015-12-25 Method and apparatus for processing high-precision map data, storage medium and device
EP15903493.3A EP3343503B1 (en) 2015-09-09 2015-12-25 High-precision map data processing method and apparatus, storage medium and device

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510571925.6A CN105260988B (zh) 2015-09-09 2015-09-09 一种高精地图数据的处理方法和装置
CN201510571925.6 2015-09-09

Publications (1)

Publication Number Publication Date
WO2017041390A1 true WO2017041390A1 (zh) 2017-03-16

Family

ID=55100663

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/099012 Ceased WO2017041390A1 (zh) 2015-09-09 2015-12-25 高精地图数据的处理方法、装置、存储介质和设备

Country Status (6)

Country Link
US (1) US10628712B2 (zh)
EP (1) EP3343503B1 (zh)
JP (1) JP6664470B2 (zh)
KR (1) KR102037820B1 (zh)
CN (1) CN105260988B (zh)
WO (1) WO2017041390A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109635052A (zh) * 2018-10-31 2019-04-16 百度在线网络技术(北京)有限公司 点云数据的处理方法、装置和存储介质
KR20190088866A (ko) * 2017-12-29 2019-07-29 바이두 온라인 네트웍 테크놀러지 (베이징) 캄파니 리미티드 포인트 클라우드 데이터 수집 궤적을 조정하기 위한 방법, 장치 및 컴퓨터 판독 가능한 매체
CN111324601A (zh) * 2020-02-09 2020-06-23 北京工业大学 一种电子地理信息系统的山上道路数据获取与生成的方法
CN114777799A (zh) * 2022-06-15 2022-07-22 中汽创智科技有限公司 一种路口高精地图生成方法、装置、电子设备及存储介质

Families Citing this family (56)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3327669B1 (en) 2016-11-26 2022-01-05 Thinkware Corporation Image processing apparatus, image processing method, computer program and computer readable recording medium
US11892311B2 (en) 2016-11-26 2024-02-06 Thinkware Corporation Image processing apparatus, image processing method, computer program and computer readable recording medium
CN108286976A (zh) * 2017-01-09 2018-07-17 北京四维图新科技股份有限公司 一种点云数据的融合方法和装置、以及混合导航系统
CN107316048B (zh) * 2017-05-03 2020-08-28 深圳市速腾聚创科技有限公司 点云分类方法及装置
CN110019603B (zh) * 2017-10-31 2025-07-22 阿里巴巴(中国)有限公司 一种道路数据融合处理方法及装置
CN109839922B (zh) * 2017-11-28 2020-11-10 百度在线网络技术(北京)有限公司 用于控制无人驾驶车辆的方法及装置
CN108030452A (zh) * 2017-11-30 2018-05-15 深圳市沃特沃德股份有限公司 视觉扫地机器人及建立场景地图的方法
CN107958451A (zh) * 2017-12-27 2018-04-24 深圳普思英察科技有限公司 视觉高精度地图制作方法和装置
CN108564657B (zh) * 2017-12-28 2021-11-16 达闼科技(北京)有限公司 一种基于云端的地图构建方法、电子设备和可读存储介质
CN108195382B (zh) * 2017-12-28 2021-09-21 湖北省测绘工程院 一种高精度导航图精度配准方法及装置
CN108460791A (zh) * 2017-12-29 2018-08-28 百度在线网络技术(北京)有限公司 用于处理点云数据的方法和装置
CN108109437B (zh) * 2018-01-24 2021-01-12 广东容祺智能科技有限公司 一种基于地图特征的无人机自主航线提取生成方法
CN110533055B (zh) * 2018-05-25 2022-04-26 北京京东尚科信息技术有限公司 一种点云数据的处理方法和装置
CN109064506B (zh) * 2018-07-04 2020-03-13 百度在线网络技术(北京)有限公司 高精度地图生成方法、装置及存储介质
CN110148144B (zh) * 2018-08-27 2024-02-13 腾讯大地通途(北京)科技有限公司 点云数据的分割方法和装置、存储介质、电子装置
CN109285220B (zh) * 2018-08-30 2022-11-15 阿波罗智能技术(北京)有限公司 一种三维场景地图的生成方法、装置、设备及存储介质
CN109146938B (zh) * 2018-08-30 2020-09-15 百度在线网络技术(北京)有限公司 动态障碍物的位置校准方法、装置、设备及存储介质
JP7136035B2 (ja) * 2018-08-31 2022-09-13 株式会社デンソー 地図生成装置及び地図生成方法
CN111061820B (zh) * 2018-10-17 2025-01-07 宝马股份公司 用于存储高精度地图的方法和装置
CN109459734B (zh) * 2018-10-30 2020-09-11 百度在线网络技术(北京)有限公司 一种激光雷达定位效果评估方法、装置、设备及存储介质
CN109597096B (zh) * 2018-12-01 2023-02-28 北醒(北京)光子科技有限公司 一种激光雷达点云处理系统及方法
CN109635701B (zh) * 2018-12-05 2023-04-18 宽凳(北京)科技有限公司 车道通行属性获取方法、装置和计算机可读存储介质
KR101996623B1 (ko) 2018-12-11 2019-07-04 주식회사 맵퍼스 Gps 궤적 품질 지표를 이용한 고정밀 지도 데이터 구축 방법 및 시스템
CN109633665A (zh) * 2018-12-17 2019-04-16 北京主线科技有限公司 交通场景稀疏激光点云拼接方法
CN109798903B (zh) * 2018-12-19 2021-03-30 广州文远知行科技有限公司 一种从地图数据中获取道路信息的方法及装置
CN109801240B (zh) * 2019-01-15 2020-12-08 武汉鸿瑞达信息技术有限公司 一种图像增强方法以及图像增强装置
CN111771229B (zh) * 2019-01-30 2023-11-21 百度时代网络技术(北京)有限公司 用于自动驾驶车辆的点云重影效果检测系统
KR102334641B1 (ko) * 2019-01-30 2021-12-03 바이두닷컴 타임즈 테크놀로지(베이징) 컴퍼니 리미티드 자율 주행 차량을 위한 맵 파티셔닝 시스템
CN111694903B (zh) * 2019-03-11 2023-09-12 北京地平线机器人技术研发有限公司 地图构建方法、装置、设备以及可读存储介质
CN109991984B (zh) * 2019-04-22 2024-04-30 上海蔚来汽车有限公司 用于生成高精细地图的方法、装置和计算机存储介质
CN110120075B (zh) * 2019-05-17 2021-09-07 百度在线网络技术(北京)有限公司 用于处理信息的方法和装置
CN110136181B (zh) * 2019-05-17 2021-08-20 百度在线网络技术(北京)有限公司 用于生成信息的方法和装置
CN110147382B (zh) * 2019-05-28 2022-04-12 北京百度网讯科技有限公司 车道线更新方法、装置、设备、系统及可读存储介质
KR102091580B1 (ko) * 2019-07-09 2020-03-20 주식회사 모빌테크 이동식 도면화 시스템을 이용한 도로 표지 정보 수집 방법
CN110516653A (zh) * 2019-09-03 2019-11-29 武汉天擎空间信息技术有限公司 一种基于多光谱机载激光雷达点云数据的道路提取方法
CN110866449A (zh) * 2019-10-21 2020-03-06 北京京东尚科信息技术有限公司 识别道路中目标对象的方法和装置
KR102288623B1 (ko) * 2019-11-18 2021-08-11 (주)이노시뮬레이션 육상이동체 시뮬레이션을 위한 정밀도로지도 데이터 처리 및 포맷변경방법
KR102276044B1 (ko) * 2019-11-25 2021-07-12 숭실대학교산학협력단 3d 포인트 클라우드의 클러스터링 방법 및 이를 수행하는 장치
CN110986969B (zh) * 2019-11-27 2021-12-28 Oppo广东移动通信有限公司 地图融合方法及装置、设备、存储介质
CN113112444B (zh) * 2020-01-09 2022-05-31 舜宇光学(浙江)研究院有限公司 鬼像检测方法及其系统、电子设备以及鬼像检测平台
CN111623788A (zh) * 2020-05-29 2020-09-04 长沙酷哇人工智能及大数据产业技术研究院有限公司 用于无人驾驶的高精地图的构建方法
CN111811530B (zh) * 2020-06-16 2022-06-21 北京五一视界数字孪生科技股份有限公司 车道线生成方法、装置、存储介质及电子设备
CN113518201B (zh) * 2020-07-14 2022-08-12 阿里巴巴集团控股有限公司 视频处理方法、装置及设备
CN112200868A (zh) * 2020-09-30 2021-01-08 深兰人工智能(深圳)有限公司 定位方法、装置和车辆
CN112537302B (zh) * 2020-11-30 2022-06-17 南通路远科技信息有限公司 一种无人驾驶交通载具车道保持方法、装置及交通载具
CN112833891B (zh) * 2020-12-31 2023-06-27 武汉光庭信息技术股份有限公司 基于卫片识别的道路数据与车道级地图数据的融合方法
CN112837241B (zh) * 2021-02-09 2025-01-10 贵州京邦达供应链科技有限公司 建图重影去除方法、设备及存储介质
CN115077537B (zh) * 2021-03-10 2025-08-01 清华大学 高精度地图感知容器设计方法及装置、存储介质、终端
CN112907746B (zh) * 2021-03-25 2024-10-29 上海商汤临港智能科技有限公司 电子地图的生成方法、装置、电子设备及存储介质
CN115407364B (zh) * 2022-09-06 2025-01-28 安徽蔚来智驾科技有限公司 点云地图处理方法、车道标注数据获取方法、设备及介质
CN115795071B (zh) * 2022-10-31 2026-03-27 毫末智行科技有限公司 一种点云数据过滤方法、装置、终端设备及存储介质
CN116320781A (zh) * 2023-02-15 2023-06-23 浙江数智交院科技股份有限公司 监控设备布设方案确定方法、装置及电子设备
CN116299534A (zh) * 2023-02-21 2023-06-23 广西柳工机械股份有限公司 车辆位姿的确定方法、装置、设备及存储介质
CN116520298B (zh) * 2023-06-12 2024-08-27 北京百度网讯科技有限公司 激光雷达性能测试方法、装置、电子设备及可读存储介质
CN117009368B (zh) * 2023-09-18 2026-01-30 奇瑞汽车股份有限公司 基于点云的高精地图更新方法、装置、车辆及介质
CN119848439B (zh) * 2025-03-20 2025-05-30 浙江华是科技股份有限公司 激光点云去噪方法及系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101673396A (zh) * 2009-09-07 2010-03-17 南京理工大学 基于动态目标检测的图像融合方法
US20100290692A1 (en) * 2007-10-11 2010-11-18 British Columbia Cancer Agency Branch Systems and methods for automated characterization of genetic heterogeneity in tissue samples
CN103996036A (zh) * 2014-06-09 2014-08-20 百度在线网络技术(北京)有限公司 一种地图数据采集方法及装置
CN104766058A (zh) * 2015-03-31 2015-07-08 百度在线网络技术(北京)有限公司 一种获取车道线的方法和装置

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7629899B2 (en) * 1997-10-22 2009-12-08 Intelligent Technologies International, Inc. Vehicular communication arrangement and method
KR0158417B1 (ko) * 1995-11-24 1999-03-20 배순훈 화상처리장치를 이용한 물체의 위치 판단방법
KR100279681B1 (ko) * 1998-05-27 2001-02-01 박권상 스포츠 영상효과 처리장치
JP2007232381A (ja) * 2006-02-27 2007-09-13 Omron Corp レーダ装置
JP5388082B2 (ja) * 2006-09-07 2014-01-15 株式会社豊田中央研究所 静止物地図生成装置
JP5286653B2 (ja) * 2006-09-07 2013-09-11 株式会社豊田中央研究所 静止物地図生成装置
JP4577580B2 (ja) 2007-04-10 2010-11-10 ソニー株式会社 位置合わせ方法、位置合わせ装置及びプログラム
JP2008276642A (ja) 2007-05-02 2008-11-13 Xanavi Informatics Corp 走行車線認識装置および走行車線認識方法
KR101657526B1 (ko) * 2012-01-11 2016-09-19 한화테크윈 주식회사 특징점 검사영역 설정기와 설정 방법 및 이를 구비하는 영상 안정화 장치
JP5883732B2 (ja) * 2012-07-03 2016-03-15 クラリオン株式会社 環境認識装置
US9424672B2 (en) * 2013-11-07 2016-08-23 Here Global B.V. Method and apparatus for processing and aligning data point clouds
CN103645480B (zh) * 2013-12-04 2015-11-18 北京理工大学 基于激光雷达和图像数据融合的地形地貌特征构建方法
US20150193971A1 (en) * 2014-01-03 2015-07-09 Motorola Mobility Llc Methods and Systems for Generating a Map including Sparse and Dense Mapping Information
US20160357262A1 (en) * 2015-06-05 2016-12-08 Arafat M.A. ANSARI Smart vehicle
US10573018B2 (en) * 2016-07-13 2020-02-25 Intel Corporation Three dimensional scene reconstruction based on contextual analysis

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100290692A1 (en) * 2007-10-11 2010-11-18 British Columbia Cancer Agency Branch Systems and methods for automated characterization of genetic heterogeneity in tissue samples
CN101673396A (zh) * 2009-09-07 2010-03-17 南京理工大学 基于动态目标检测的图像融合方法
CN103996036A (zh) * 2014-06-09 2014-08-20 百度在线网络技术(北京)有限公司 一种地图数据采集方法及装置
CN104766058A (zh) * 2015-03-31 2015-07-08 百度在线网络技术(北京)有限公司 一种获取车道线的方法和装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3343503A4 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20190088866A (ko) * 2017-12-29 2019-07-29 바이두 온라인 네트웍 테크놀러지 (베이징) 캄파니 리미티드 포인트 클라우드 데이터 수집 궤적을 조정하기 위한 방법, 장치 및 컴퓨터 판독 가능한 매체
KR102068419B1 (ko) * 2017-12-29 2020-01-21 바이두 온라인 네트웍 테크놀러지 (베이징) 캄파니 리미티드 포인트 클라우드 데이터 수집 궤적을 조정하기 위한 방법, 장치 및 컴퓨터 판독 가능한 매체
CN109635052A (zh) * 2018-10-31 2019-04-16 百度在线网络技术(北京)有限公司 点云数据的处理方法、装置和存储介质
CN111324601A (zh) * 2020-02-09 2020-06-23 北京工业大学 一种电子地理信息系统的山上道路数据获取与生成的方法
CN114777799A (zh) * 2022-06-15 2022-07-22 中汽创智科技有限公司 一种路口高精地图生成方法、装置、电子设备及存储介质
CN114777799B (zh) * 2022-06-15 2022-10-18 中汽创智科技有限公司 一种路口高精地图生成方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
US10628712B2 (en) 2020-04-21
KR20180041176A (ko) 2018-04-23
US20180253625A1 (en) 2018-09-06
JP2018534603A (ja) 2018-11-22
CN105260988B (zh) 2019-04-05
JP6664470B2 (ja) 2020-03-13
EP3343503A4 (en) 2018-08-08
EP3343503B1 (en) 2020-02-19
CN105260988A (zh) 2016-01-20
KR102037820B1 (ko) 2019-10-29
EP3343503A1 (en) 2018-07-04

Similar Documents

Publication Publication Date Title
KR102037820B1 (ko) 고정밀 지도 데이터 처리 방법, 장치, 저장 매체 및 기기
CN112380317B (zh) 高精地图更新方法、装置、电子设备和存储介质
CN112949366B (zh) 障碍物识别方法和装置
JP6561199B2 (ja) レーザ点群に基づく都市道路の認識方法、装置、記憶媒体及び機器
CN109061703B (zh) 用于定位的方法、装置、设备和计算机可读存储介质
KR102145109B1 (ko) 지도 생성 및 운동 객체 위치 결정 방법 및 장치
US10074020B2 (en) Vehicular lane line data processing method, apparatus, storage medium, and device
US20190147320A1 (en) "Matching Adversarial Networks"
WO2018068653A1 (zh) 点云数据处理方法、装置及存储介质
CN110674705B (zh) 基于多线激光雷达的小型障碍物检测方法及装置
US9658074B2 (en) Diverging and converging road geometry generation from sparse data
WO2021017211A1 (zh) 一种基于视觉的车辆定位方法、装置及车载终端
CN114485698B (zh) 一种交叉路口引导线生成方法及系统
CN108345822A (zh) 一种点云数据处理方法及装置
CN110210384B (zh) 一种道路全局信息实时提取与表示系统
CN117576652B (zh) 道路对象的识别方法、装置和存储介质及电子设备
CN112651266B (zh) 行人检测方法和装置
CN115139303A (zh) 一种栅格井盖检测方法、装置、设备及存储介质
US11461944B2 (en) Region clipping method and recording medium storing region clipping program
CN115909262A (zh) 车位检测方法、装置、电子设备以及存储介质
WO2023088486A1 (zh) 车道线提取方法、装置、车辆及存储介质
CN116412804A (zh) 交通标志牌的激光雷达视觉融合建图方法、装置及介质
CN109144052A (zh) 用于自动驾驶车辆的导航系统及其方法
CN118628767A (zh) 一种基于点云地图的动态目标点云剔除方法、装置、电子设备及存储介质
CN117392216A (zh) 点云地图的确定方法、装置、电子设备及存储介质

Legal Events

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

Ref document number: 15903493

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 15758623

Country of ref document: US

ENP Entry into the national phase

Ref document number: 2018512859

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 20187007384

Country of ref document: KR

Kind code of ref document: A

WWE Wipo information: entry into national phase

Ref document number: 2015903493

Country of ref document: EP