WO2020062856A1 - 一种车辆特征获取方法及装置 - Google Patents

一种车辆特征获取方法及装置 Download PDF

Info

Publication number
WO2020062856A1
WO2020062856A1 PCT/CN2019/084537 CN2019084537W WO2020062856A1 WO 2020062856 A1 WO2020062856 A1 WO 2020062856A1 CN 2019084537 W CN2019084537 W CN 2019084537W WO 2020062856 A1 WO2020062856 A1 WO 2020062856A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
image
area
processed
sideways
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/CN2019/084537
Other languages
English (en)
French (fr)
Inventor
田欢
胡骏
程帅
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.)
Neusoft Reach Automotive Technology Shenyang Co Ltd
Original Assignee
Neusoft Reach Automotive Technology Shenyang 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 Neusoft Reach Automotive Technology Shenyang Co Ltd filed Critical Neusoft Reach Automotive Technology Shenyang Co Ltd
Priority to US17/280,191 priority Critical patent/US12002269B2/en
Priority to EP19864666.3A priority patent/EP3859593B1/en
Priority to JP2021542238A priority patent/JP7190583B2/ja
Publication of WO2020062856A1 publication Critical patent/WO2020062856A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • 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
    • G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30248—Vehicle exterior or interior
    • G06T2207/30252—Vehicle exterior; Vicinity of vehicle
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
    • G06V10/422—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation for representing the structure of the pattern or shape of an object therefor
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • 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
    • G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
    • G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/08—Detecting or categorising vehicles

Definitions

  • the present invention relates to the field of automobiles, and in particular, to a method and device for acquiring vehicle characteristics.
  • the target image in the target image can be obtained by analyzing the target image to obtain the characteristics of the target object, so that subsequent analysis can be performed according to the characteristics of the target object. For example, by identifying target objects such as vehicles, buildings, and pedestrians on the road in the target image, the color, shape, size, and position characteristics of these target objects can be obtained, and road conditions can be determined based on these characteristics.
  • the target object in the target image will be reflected as a flat area.
  • the vehicles and pedestrians in the target image can be reflected as areas within a rectangular frame.
  • the size of this area is related to the size of the target object and also
  • the shooting parameters of the target image are related, and the color of the pixels in the area is related to the color of the target object.
  • embodiments of the present application provide a method and device for acquiring vehicle features.
  • the vehicle in the image to be processed is divided into regions, and then the regional features are obtained in the divided regions, so that a more comprehensive vehicle can be obtained. feature.
  • An embodiment of the present application provides a method for acquiring vehicle characteristics, where the method includes:
  • the characteristic element includes a side body element and a vehicle end element, and the vehicle end is a front or rear end of the vehicle;
  • the determining the sideways area and the area of the vehicle end according to the positions of the sideways elements and the vehicle end elements in the image to be processed includes:
  • the method further includes:
  • the image shooting device Determining the relative position of the vehicle and an image capture device according to the area characteristics of the vehicle end and the area characteristics of the sideways, and the shooting parameters of the image to be processed, the image shooting device being configured to shoot the vehicle The image to be processed is obtained.
  • determining the relative position of the vehicle and the image capturing device according to the area characteristics of the vehicle end and the area characteristics of the sideways body and the shooting parameters of the image to be processed includes:
  • determining the relative position of the vehicle and the image acquisition device according to the area characteristics of the vehicle end and the area characteristics of the sideways, and the shooting parameters of the image to be processed includes:
  • a relative position of the vehicle and the image capturing device is determined according to a boundary between the region of the vehicle end and the sideways region, and shooting parameters of the image to be processed.
  • the characteristics of the sideways area include the position of the apex of the sideways area, the center point of the sideways area, or the midpoint of the side of the sideways area in the image to be processed; the area of the car end
  • the features include the position of a vertex of the region of the vehicle end, a center point of the region of the vehicle end, or a midpoint of an edge of the region of the vehicle end in the image to be processed.
  • the identifying the image to be processed to obtain a characteristic element on the vehicle includes:
  • the elements of the front include one or more of a front lamp, a front window, an insurance handle, and a front license plate
  • the elements of the rear include a rear lamp, a rear window, and a rear license plate
  • the sideways elements include one or more of a wheel, a side window, a rearview mirror, and a door.
  • An embodiment of the present application provides a vehicle feature acquisition device, where the device includes:
  • An image acquisition unit configured to acquire an image to be processed, where the image to be processed includes a vehicle
  • An image recognition unit configured to identify the image to be processed to obtain characteristic elements on the vehicle;
  • the characteristic elements include side elements and vehicle end elements, and the vehicle end is the front or rear of the vehicle;
  • An area determining unit configured to determine the sideways area and the area of the vehicle end of the vehicle according to the positions of the sideways element and the side of the vehicle element in the image to be processed;
  • a feature acquiring unit is configured to acquire a sideways area feature according to the sideways area, and obtain a vehicle side area feature according to the vehicle side area.
  • the area determining unit includes:
  • a vehicle area determining unit configured to determine a vehicle area in which the vehicle is located in the image to be processed
  • a dividing line determining unit configured to determine the division of the sideways area and the area of the car end according to the position of the sideways element near the side end element and the sideways element near the side end element in the image to be processed;
  • a region determining subunit configured to determine, based on the dividing line, and the positions of the sideways elements and the vehicle end elements in the image to be processed, the sideways area and the vehicle end in the vehicle region; region.
  • the device further includes:
  • a relative position determining unit configured to determine a relative position of the vehicle and an image capturing device according to the area characteristics of the vehicle end and the area characteristics of the sideways, and the shooting parameters of the image to be processed;
  • the image to be processed is obtained by shooting the vehicle.
  • the relative position determining unit includes:
  • a target area determining unit configured to determine an area of the vehicle end and the sideways area that faces the image capturing device as a target area
  • the relative position determining subunit is configured to determine a relative position of the vehicle and the image capturing device according to a region feature of the target region and a shooting parameter of the image to be processed.
  • the relative position determining unit is specifically configured to:
  • a relative position of the vehicle and the image capturing device is determined according to a boundary between the region of the vehicle end and the sideways region, and shooting parameters of the image to be processed.
  • the characteristics of the sideways area include the position of the apex of the sideways area, the center point of the sideways area, or the midpoint of the side of the sideways area in the image to be processed; the area of the car end
  • the features include the position of a vertex of the region of the vehicle end, a center point of the region of the vehicle end, or a midpoint of an edge of the region of the vehicle end in the image to be processed.
  • the image recognition unit includes:
  • a vehicle area identification unit configured to identify in the image to be processed and determine a vehicle area where the vehicle is located
  • a vehicle image acquisition unit configured to intercept the vehicle region in the image to be processed to form a vehicle image
  • a feature element recognition unit is configured to identify the vehicle image and determine a feature element on the vehicle.
  • the elements of the front include one or more of a front lamp, a front window, an insurance handle, and a front license plate
  • the elements of the rear include a rear lamp, a rear window, and a rear license plate
  • the sideways elements include one or more of a wheel, a side window, a rearview mirror, and a door.
  • an image to be processed is acquired.
  • the image to be processed includes a vehicle, and the image to be processed is identified to obtain a characteristic element on the vehicle.
  • the characteristic element may be an element of a side body of the vehicle and an element of a vehicle end.
  • the vehicle end may be The front or rear of the vehicle can determine the side of the vehicle and the area of the front of the vehicle based on the position of the side of the vehicle and the elements of the end of the vehicle in the image to be processed, and then obtain the characteristics of the side of the vehicle based on the side of the vehicle.
  • Regional characteristics Therefore, in the implementation of the present application, the vehicle in the image to be processed is divided into regions, and then the regional features are obtained in the divided region. Compared with the prior art, only the planar region of the vehicle in the image to be processed is determined. Can get more comprehensive vehicle characteristics.
  • FIG. 1 is a schematic diagram of acquiring features through a target image in the prior art
  • FIG. 2 is a flowchart of a method for acquiring vehicle characteristics according to an embodiment of the present application
  • FIG. 3 is a schematic diagram of a feature element recognition process provided by an embodiment of the present application.
  • FIG. 4 is a schematic diagram of an area determination process according to an embodiment of the present application.
  • FIG. 5 is a structural block diagram of a vehicle feature acquisition device according to an embodiment of the present application.
  • the target image in the target image can be obtained by analyzing the target image, obtaining the characteristics of the target object, and performing subsequent analysis according to the characteristics of the target object.
  • the color, shape, size, and position of these target objects can be obtained from target objects such as vehicles, buildings, and pedestrians on the road in the target image, and road conditions can be determined based on these characteristics.
  • the target object in the target image will be reflected as a flat area.
  • FIG. 1 for a schematic diagram of acquiring features through the target image in the prior art.
  • the vehicles and pedestrians in the target image are respectively represented as The area within the rectangular frame of different colors, taking the vehicle as an example, the size of the vehicle area is related to the size of the vehicle and also to the shooting parameters of the vehicle.
  • the color of the pixel points in the vehicle area is related to the color of the vehicle.
  • the vehicle has certain three-dimensional characteristics, and the planar area does not well reflect the three-dimensional characteristics of the vehicle, such as the placement form of the vehicle and the distance from the image capturing equipment.
  • An image capturing device is a device that captures a vehicle to obtain a target image.
  • FIG. 1 a schematic diagram of acquiring features by using a target image in the prior art is shown.
  • the vehicle in the lower left corner is closer to the left rear end of the image capturing device.
  • Center to calculate the distance between the vehicle and the image-capturing equipment, then the calculated distance should be the distance between the vehicle's interior and the image-capturing equipment, so the calculated distance should be greater than the actual distance between the vehicle and the image-capturing equipment.
  • Assisted driving may lead to more serious consequences.
  • an image to be processed may be acquired.
  • the image to be processed includes a vehicle, and the image to be processed is identified to obtain a characteristic element on the vehicle, where the characteristic element may be an element on the side of the vehicle and an element on the end of the vehicle.
  • the vehicle end can be the front or rear of the vehicle.
  • the sideways area and the area of the vehicle end can be determined, and the sideways area characteristics can be obtained according to the sideways area.
  • the area at the end obtains the area characteristics of the car end.
  • the vehicle in the image to be processed is divided into regions, and then the regional features are obtained in the divided region. Compared with the prior art, only the position of the vehicle in the image to be processed is determined. , Can get more comprehensive vehicle characteristics.
  • the method may include the following steps.
  • the target image is used as the image to be processed.
  • the image to be processed is an image including a vehicle, and can be captured by an image capturing device.
  • the image capturing device may be a camera, a video camera, a camera, or the like.
  • the image to be processed may include only a vehicle or other target objects other than the vehicle.
  • the image to be processed may include one vehicle or multiple vehicles.
  • the characteristic elements on the vehicle may be sideways elements and elements of the vehicle end, where the vehicle end may be the front or rear of the vehicle.
  • the front element is an element that can assist in identifying the front position.
  • the rear element is an element that can assist in identifying the rear position.
  • it can include one or more of rear lights, rear windows, and rear license plates.
  • Sideways elements are elements that can help identify the sideways position. For example, they can be wheels, side windows, rearview mirrors, and doors.
  • Identifying the image to be processed to obtain the characteristic elements on the vehicle may be specifically: segmenting the image to be processed to obtain the category labels of each pixel in the image to be processed. Among the segmented images of the image to be processed, the same feature element The category labels of the pixels are the same, and the feature elements on the vehicle are obtained according to the category labels. Segmentation of the image to be processed can be performed by a deep learning neural network trained in advance. Specifically, the red oval area can be identified as the headlight area, the square large gray area is the window area, and the circle has a white ray shape. The areas are wheels and so on.
  • different colors can be preset for different category labels to form a segmented image.
  • the color corresponding to the category label of the lamp area can be red
  • the window The color corresponding to the category label of the area may be blue or the like.
  • the recognition of feature elements in the image to be processed can be divided into two steps: identifying the image to be processed and determining the vehicle where the vehicle is located Area, intercept the vehicle area in the image to be processed to form a target image, identify the target image, and determine the characteristic elements in the vehicle area.
  • the first deep learning neural network obtained through pre-training can be used to identify the image to be processed to obtain a first segmented image, where each pixel in the first segmented image has a category label and the category labels of pixels of the same target object The same, so that the vehicle area in the image to be processed is identified according to the category label;
  • the target image can be identified through a second deep learning neural network trained in advance to obtain a second segmented image, where each pixel in the second segmented image has Category labels, and the category labels of the same feature element are the same.
  • the category labels the feature elements in the vehicle area can be identified.
  • FIG. 3 a schematic diagram of a feature element recognition process provided in real time in this application is shown in FIG. 3, where FIG. 3 (a) is an exemplary image to be processed, which is processed by a first deep learning neural network. Segmentation to obtain the first segmented image shown in FIG. 3 (b), where the area in the rectangular frame is the vehicle area where the vehicle is located; the rectangular frame area in the first segmented image is enlarged to obtain a reference to FIG. 3 (c) The vehicle segmentation image shown; according to the vehicle segmentation image, the image to be processed is intercepted to obtain the vehicle image shown in FIG. 3 (d); the vehicle image is segmented by the second deep learning neural network to obtain the image shown in FIG. 3 (e). The second segmented image shown, where two circular areas are the lights on the vehicle, and a rectangular area is the windows on the vehicle.
  • a tire or a rearview mirror can also be identified. And other elements.
  • S103 Determine the sideways area and the area of the vehicle end according to the positions of the sideways element and the side of the vehicle element in the image to be processed.
  • the vehicle in the image to be processed includes at least sideways elements and elements of the vehicle end, and the sideways region and the region of the vehicle end may be directly determined according to the sideways elements and the elements of the vehicle end.
  • elements at the rear of the vehicle such as rear lights, windows, etc.
  • sideways elements such as wheels, rearview mirrors, etc.
  • the wheels in the sideways elements are rear Side wheels, thereby determining that the region of the rear of the image to be processed is a rectangular region including the vehicle, and the side of the region is the left and rear edges of the region of the rear of the vehicle.
  • the vehicle area where the vehicle is located, and the boundary between the sideways area and the vehicle end area may be determined, and then determined in the vehicle area according to the boundary line. Sideways and end-of-car areas.
  • FIG. 4 is a schematic diagram of an area determination process according to an embodiment of the present application.
  • the image shown in FIG. 4 (b) can be obtained by image segmentation.
  • Vehicle area gray rectangular frame
  • image segmentation of the image to be processed can obtain the characteristic elements shown in FIG.
  • the front end is the front
  • the sideways elements of the elements near the front are the right wheel in the image to be processed, and the right edge of the side window in the image to be processed
  • the elements near the front of the side element are the left of the front window. Edge, and the left headlight of the front, so you can determine the front and sideways areas based on the right wheel in the image to be processed, the right edge of the side window, the left edge of the front window, and the left headlight of the front Dividing line.
  • the boundary between the front area and the side area can be determined in various ways, for example, according to the left edge of the front window and the right edge of the side window, or the right edge of the right wheel and the left of the left light.
  • the edges are OK.
  • the dividing line can be vertical or inclined. In the embodiment of the present application, as an example, a vertical straight line may be made through the right edge of the right wheel, as the boundary between the front area and the sideways area. Refer to the gray line shown in FIG. 4 (d).
  • the sideline area and the vehicle end area may be determined in the vehicle area according to the determined boundary line and the positions of the sideside elements and the vehicle end elements in the image to be processed. It can be understood that the sideways area and the vehicle end area are located on both sides of the dividing line, and the sideways area and the vehicle end area can be rectangular areas, parallelogram areas, or other shapes that can represent areas.
  • Analogous to the determination of the dividing line there are multiple ways to determine the sideways area and the vehicle end area in the vehicle area.
  • the vertical straight line passing the right edge of the right wheel is used as the front area and sideways area.
  • the dividing line as shown in FIG. 4 (e), can use the horizontal line passing the contact point between the wheel and the ground as the lower edge of the front area, the dividing line as the left edge of the front area, and the horizontal line of the roof as the front.
  • the top edge of the area uses the right edge of the vehicle area as the right edge of the front area.
  • the four edge lines form the front area (a rectangular frame made of white thick lines on the right side); similarly, the dividing line is used as the right edge of the side area.
  • the sideways area and the vehicle end area can be represented by the coordinates of the six vertices in the image to be processed. Since there can be multiple ways to determine the boundary, the sideways area and the vehicle side area can also be determined in multiple ways. Therefore, multiple sets of vertex coordinate values can be obtained, and weighted averages of multiple sets of vertex coordinate values can be used to obtain weighted vertex coordinate values that represent the sideways area and the vehicle end area to improve the accuracy of the sideways area and the vehicle end area.
  • the sideways area feature may be the position of the feature point such as the apex of the sideways area, the center point of the sideways area, or the midpoint of the side of the sideways area in the image to be processed.
  • the area feature of the vehicle end includes the area of the vehicle end.
  • the position of the feature point in the image to be processed such as the vertex of the vehicle, the center of the region of the vehicle end, or the center of the edge of the region of the vehicle end.
  • the regional feature of the vehicle head may be the vertex of the vehicle head region, the center point of the vehicle head region, or the vehicle head.
  • the position of feature points such as the midpoint of the edges of the region in the image to be processed. It can be understood that the position of the feature point in the image to be processed can be represented by the coordinates of the feature point in the image to be processed.
  • the 3D feature area of the vehicle can be obtained based on the vertices of the side features and the vertices of the area features.
  • the 3D feature areas of the vehicle can include the front area and the rear area. , Sideways area on both sides, roof area, and bottom area.
  • an image to be processed is first obtained.
  • the image to be processed includes a vehicle, and the image to be processed is identified to obtain a characteristic element on the vehicle.
  • the characteristic element may be a sideways element of the vehicle.
  • the end of the vehicle, the front of the vehicle can be the front or the rear of the vehicle, and then based on the position of the side of the vehicle and the position of the end of the vehicle in the image to be processed, the side of the vehicle and the area of the vehicle can be determined, and the side can be obtained according to the side
  • the regional characteristics of the vehicle are obtained based on the regional characteristics of the vehicle. Therefore, the vehicle in the image to be processed is divided into regions, and then the regional features are obtained in the divided region. Compared with the prior art, which only determines the position of the vehicle in the image to be processed, a more comprehensive vehicle can be obtained. feature.
  • the vehicle characteristics are obtained to perform corresponding analysis through the vehicle characteristics to obtain relevant results.
  • the embodiments of the present application may also be based on the vehicle side characteristics.
  • the area feature and the side area feature, as well as the shooting parameters of the image to be processed, determine the relative position of the vehicle and the image capturing device.
  • the image capture device is a device that captures a vehicle to obtain an image to be processed.
  • the shooting parameters of the image to be processed refer to parameters such as the focal length, wide angle, position, and rotation angle when the image capture device captures the vehicle.
  • the relative position of the feature point and the image capturing device can be determined according to the position of the feature point in the image to be processed and the shooting parameters of the image to be processed. Reflects the three-dimensional characteristics of the vehicle, so the relative position of the vehicle and the image capturing device can be obtained accordingly.
  • determining the relative position of the vehicle and the image capturing device according to the area characteristics of the vehicle side and the area characteristics of the side body, and the shooting parameters of the image to be processed may be specifically, combining the area of the vehicle side and the side body area, The area facing the image-capturing device is determined as the target area, and the relative position of the vehicle and the image-capturing device is determined according to the area characteristics of the target area and the shooting parameters of the image to be processed.
  • the target area facing the image capturing device may be determined according to the area and / or shape of the vehicle-end area and the sideways area, for example, a larger area is determined as the target area facing the image capturing device, or a rectangular shape area is determined Towards the target area of the image capture device.
  • the area of the front of the vehicle is larger than the area of the sideways area, it can be considered that the area of the front of the vehicle faces the image capture device, so the area of the front of the vehicle can be determined as the target area; Since the shape of the sideways area is a parallelogram, it can be considered that the area of the front of the vehicle faces the image capturing device, so the area of the front of the vehicle can be determined as the target area.
  • the regional characteristics of the target region may be the positions of the characteristic points such as the apex of the target region, the center point of the target region, or the midpoint of the edges of the target region in the image to be processed.
  • the relative positions of these feature points and the image capturing device can be determined, and based on this, the relative positions of the vehicle and the image capturing device can be obtained.
  • the target area is the front area.
  • the front area can be determined according to the position of the center point of the front area in the image to be processed and the shooting parameters of the image to be processed. The relative position of the center point of the image capture device.
  • determining the relative position of the vehicle and the image capturing device according to the area characteristics of the vehicle side and the area characteristics of the side body and the shooting parameters of the image to be processed may be specifically based on the area of the vehicle side and the side body area.
  • the dividing line and the shooting parameters of the image to be processed determine the relative position of the vehicle and the image capturing device.
  • the midpoint of the dividing line may be determined, and the relative position of the vehicle and the image capturing device may be determined according to the midpoint of the dividing line and the shooting parameters of the image to be processed. This is because the boundary between the vehicle end area and the sideways area is often the closest to the image capture device, so the relative position of the vehicle and the image capture device is relatively accurate.
  • an embodiment of the present application further provides a vehicle characteristic acquiring device.
  • a vehicle characteristic acquiring device provided in an embodiment of the present application.
  • the device may include:
  • An image acquisition unit 110 configured to acquire an image to be processed, where the image to be processed includes a vehicle;
  • An image recognition unit 120 is configured to identify the image to be processed to obtain a characteristic element on the vehicle;
  • the characteristic element includes a side element and a vehicle end element, and the vehicle end is a front or rear end of the vehicle;
  • An area determining unit 130 configured to determine the sideways area and the area of the vehicle end of the vehicle according to the positions of the sideways element and the side of the vehicle element in the image to be processed;
  • a feature obtaining unit 140 is configured to obtain the side features of the side body according to the side body area, and obtain the area features of the car side according to the area of the car side.
  • the area determining unit includes:
  • a vehicle area determining unit configured to determine a vehicle area in which the vehicle is located in the image to be processed
  • a dividing line determining unit configured to determine a dividing line between the sideways area and the area of the car end according to the position of the sideways element near the side of the vehicle element and the sideways element near the side of the vehicle element in the image to be processed;
  • An area determining subunit is configured to determine a sideways area and a vehicle end area in the vehicle area according to the boundary line and positions of the sideways element and the vehicle end element in the image to be processed.
  • the device further includes:
  • the relative position determining unit is configured to determine a relative position of the vehicle and the image capturing device according to the area characteristics of the vehicle end and the area characteristics of the sideways, and the shooting parameters of the image to be processed.
  • the relative position determining unit includes:
  • a target area determining unit configured to determine, from the area at the vehicle end and the sideways area, an area facing an image capturing device as a target area, where the image capturing device is configured to capture the vehicle to obtain a to-be-processed image;
  • a relative position determining subunit is configured to determine a relative position of the vehicle and the image capturing device according to a regional feature of a target region and a shooting parameter of the image to be processed.
  • the relative position determining unit is specifically configured to:
  • the characteristics of the sideways area include the position of the apex of the sideways area, the center point of the sideways area, or the midpoint of the side of the sideways area in the image to be processed; the area of the car end
  • the features include the position of a vertex of the region of the vehicle end, a center point of the region of the vehicle end, or a midpoint of an edge of the region of the vehicle end in the image to be processed.
  • the image recognition unit includes:
  • a vehicle area identification unit configured to identify in the image to be processed and determine a vehicle area where the vehicle is located
  • a vehicle image acquisition unit configured to intercept the vehicle region in the image to be processed to form a vehicle image
  • a characteristic element recognition unit is configured to identify the vehicle image and determine a characteristic element on the vehicle.
  • the elements of the front include one or more of a front lamp, a front window, an insurance handle, and a front license plate
  • the elements of the rear include a rear lamp, a rear window, and a rear license plate
  • the sideways elements include one or more of a wheel, a side window, a rearview mirror, and a door.
  • an image to be processed is first acquired, the image to be processed includes a vehicle, and the image to be processed is identified to obtain a characteristic element on the vehicle, where the characteristic element may be a sideways element of the vehicle And the end of the vehicle, the front of the vehicle can be the front or the rear of the vehicle, and then based on the position of the side of the vehicle and the position of the end of the vehicle in the image to be processed, the side of the vehicle and the area of the vehicle can be determined, and the side can be obtained according to the side of the vehicle.
  • the regional characteristics of the vehicle are obtained based on the regional characteristics of the vehicle. Therefore, the vehicle in the image to be processed is divided into regions, and then the regional features are obtained in the divided region. Compared with the prior art, which only determines the position of the vehicle in the image to be processed, a more comprehensive vehicle can be obtained. feature.
  • the technical solution of the present invention can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, such as a read-only memory (English: read-only memory (ROM) / RAM, magnetic disk,
  • ROM read-only memory
  • the optical disk and the like include a plurality of instructions for causing a computer device (which may be a personal computer, a server, or a network communication device such as a router) to perform the method described in each embodiment of the present invention or some parts of the embodiment.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

本发明实施例公开了一种车辆特征获取方法及装置，获取待处理图像，待处理图像中包括车辆，对待处理图像进行识别，得到车辆上的特征元素，其中特征元素可以是车辆的侧身元素和车端的元素，车端可以是车头或车尾，根据车辆的侧身元素和车端的元素在待处理图像中的位置，可以确定车辆的侧身区域和车端的区域，进而可以根据侧身区域获取侧身的区域特征，根据车端的区域获取车端的区域特征。因此，本申请实施中，先对待处理图像中的车辆进行了区域的划分，再在划分得到的区域中获取区域特征，相比于现有技术中只确定车辆在待处理图像中的平面区域，能够得到更全面的车辆特征。

Description

一种车辆特征获取方法及装置
本申请要求于2018年09月27日提交中国专利局、申请号为201811132325.X、申请名称为“一种车辆特征获取方法及装置”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
本发明涉及汽车领域，特别是涉及一种车辆特征获取方法及装置。
背景技术
现有技术中，可以通过对目标图像进行分析，识别得到目标图像中的目标对象，获取目标对象的特征，从而可以根据目标对象的特征进行后续分析。举例来说，通过对目标图像中道路上的车辆、建筑及行人等目标对象进行识别，可以得到这些目标对象的颜色、形状、大小以及位置等特征，根据这些特征可以确定道路路况。
通常情况下，目标图像中的目标对象会被体现为一个平面的区域，目标图像中的车辆和行人可以分别被体现为矩形框内的区域，该区域的大小与目标对象的大小相关，也与目标图像的拍摄参数相关，该区域中的像素点的颜色与目标对象的颜色相关。
然而，在实际应用中，仅仅确定车辆在目标图像中所在的平面区域，往往是不全面的。
发明内容
为解决上述技术问题，本申请实施例提供一种车辆特征获取方法和装置，对待处理图像中的车辆进行了区域的划分，再在划分得到的区域中获取区域特征，因此能够得到较全面的车辆特征。
本申请实施例提供一种车辆特征获取方法，所述方法包括：
获取待处理图像，所述待处理图像中包括车辆；
对所述待处理图像进行识别，得到所述车辆上的特征元素；所述特征元素包括侧身元素和车端的元素，所述车端为车头或车尾；
根据所述侧身元素和所述车端的元素在所述待处理图像中的位置，确定所 述车辆的侧身区域和车端的区域；
根据所述侧身区域获取侧身的区域特征，根据所述车端的区域获取车端的区域特征。
可选的，所述根据所述侧身元素和所述车端的元素在所述待处理图像中的位置，确定所述车辆的侧身区域和车端的区域，包括：
在所述待处理图像中确定所述车辆所在的车辆区域；
根据靠近所述侧身元素的车端的元素和靠近所述车端的元素的侧身元素在所述待处理图像中的位置，确定所述侧身区域和所述车端的区域的分界线；
根据所述分界线，以及所述侧身元素和所述车端的元素在所述待处理图像中的位置，在所述车辆区域中确定所述侧身区域和所述车端的区域。
可选的，所述方法还包括：
根据所述车端的区域特征和所述侧身的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与图像拍摄设备的相对位置，所述图像拍摄设备用于对所述车辆进行拍摄得到所述待处理图像。
可选的，所述根据所述车端的区域特征和所述侧身的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与图像拍摄设备的相对位置，包括：
将所述车端的区域和所述侧身区域中，朝向所述图像拍摄设备的区域确定为目标区域；
根据所述目标区域的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像拍摄设备的相对位置。
可选的，所述根据所述车端的区域特征和所述侧身的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像获取设备的相对位置，包括：
根据所述车端的区域和所述侧身区域的分界线，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像拍摄设备的相对位置。
可选的，所述侧身的区域特征包括所述侧身区域的顶点、所述侧身区域的中心点或所述侧身区域的边的中点在所述待处理图像中的位置；所述车端的区域特征包括所述车端的区域的顶点、所述车端的区域的中心点或所述车端的区域的边的中点在所述待处理图像中的位置。
可选的，所述对所述待处理图像进行识别，得到所述车辆上的特征元素，包括：
在所述待处理图像中进行识别，确定所述车辆所在的车辆区域；
在所述待处理图像中截取所述车辆区域形成车辆图像；
对所述车辆图像进行识别，确定所述车辆上的特征元素。
可选的，所述车头的元素包括前车灯、前车窗、保险扛和前车牌中的一项或多项，所述车尾的元素包括后车灯、后车窗和后车牌中的一项或多项，所述侧身元素包括车轮、侧车窗、后视镜和车门中的一项或多项。
本申请实施例提供一种车辆特征获取装置，所述装置包括：
图像获取单元，用于获取待处理图像，所述待处理图像中包括车辆；
图像识别单元，用于对所述待处理图像进行识别，得到所述车辆上的特征元素；所述特征元素包括侧身元素和车端的元素，所述车端为车头或车尾；
区域确定单元，用于根据所述侧身元素和所述车端的元素在所述待处理图像中的位置，确定所述车辆的侧身区域和车端的区域；
特征获取单元，用于根据所述侧身区域获取侧身的区域特征，根据所述车端的区域获取车端的区域特征。
可选的，所述区域确定单元，包括：
车辆区域确定单元，用于在所述待处理图像中确定所述车辆所在的车辆区域；
分界线确定单元，用于根据靠近所述侧身元素的车端的元素和靠近所述车端的元素的侧身元素在所述待处理图像中的位置，确定所述侧身区域和所述车端的区域的分界线；
区域确定子单元，用于根据所述分界线，以及所述侧身元素和所述车端的元素在所述待处理图像中的位置，在所述车辆区域中确定所述侧身区域和所述车端的区域。
可选的，所述装置还包括：
相对位置确定单元，用于根据所述车端的区域特征和所述侧身的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与图像拍摄设备的相对位置，所述图像拍摄设备用于对所述车辆进行拍摄得到所述待处理图像。
可选的，所述相对位置确定单元，包括：
目标区域确定单元，用于将所述车端的区域和所述侧身区域中，朝向图像拍摄设备的区域确定为目标区域；
相对位置确定子单元，用于根据所述目标区域的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像拍摄设备的相对位置。
可选的，所述相对位置确定单元具体用于：
根据所述车端的区域和所述侧身区域的分界线，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像拍摄设备的相对位置。
可选的，所述侧身的区域特征包括所述侧身区域的顶点、所述侧身区域的中心点或所述侧身区域的边的中点在所述待处理图像中的位置；所述车端的区域特征包括所述车端的区域的顶点、所述车端的区域的中心点或所述车端的区域的边的中点在所述待处理图像中的位置。
可选的，所述图像识别单元，包括：
车辆区域识别单元，用于在所述待处理图像中进行识别，确定所述车辆所在的车辆区域；
车辆图像获取单元，用于在所述待处理图像中截取所述车辆区域形成车辆图像；
特征元素识别单元，用于对所述车辆图像进行识别，确定所述车辆上的特征元素。
可选的，所述车头的元素包括前车灯、前车窗、保险扛和前车牌中的一项或多项，所述车尾的元素包括后车灯、后车窗和后车牌中的一项或多项，所述侧身元素包括车轮、侧车窗、后视镜和车门中的一项或多项。
在本申请实施例中，获取待处理图像，待处理图像中包括车辆，对待处理图像进行识别，得到车辆上的特征元素，其中特征元素可以是车辆的侧身元素和车端的元素，车端可以是车头或车尾，根据车辆的侧身元素和车端的元素在待处理图像中的位置，可以确定车辆的侧身区域和车端的区域，进而可以根据侧身区域获取侧身的区域特征，根据车端的区域获取车端的区域特征。因此，本申请实施中，先对待处理图像中的车辆进行了区域的划分，再在划分得到的区域中获取区域特征，相比于现有技术中只确定车辆在待处理图像中的平面区域，能够得到更全面的车辆特征。
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中记载的一些实施例，对于本领域普通技术人员来讲，还可以根据这些附图获 得其他的附图。
图1为现有技术中通过目标图像获取特征的示意图；
图2为本申请实施例提供的一种车辆特征获取方法的流程图；
图3为本申请实施例提供的一种特征元素识别过程示意图；
图4为本申请实施例提供的一种区域确定过程示意图；
图5为本申请实施例提供的一种车辆特征获取装置的结构框图。
具体实施方式
发明人经过研究发现，现有技术中，可以通过对目标图像进行分析，识别得到目标图像中的目标对象，获取目标对象的特征，从而根据目标对象的特征进行后续分析。举例来说，通过对目标图像中道路上的车辆、建筑及行人等目标对象，可以得到这些目标对象的颜色、形状、大小以及位置等特征，根据这些特征可以确定道路路况。
通常情况下，目标图像中的目标对象会被体现为一个平面的区域，参考图1所示为现有技术中通过目标图像获取特征的示意图，其中，目标图像中的车辆和行人分别被体现为不同颜色的矩形框内的区域，以车辆为例，车辆区域的大小与车辆的大小相关，也与车辆的拍摄参数相关，车辆区域中的像素点的颜色与车辆的颜色相关。
然而，在实际应用中，仅仅确定车辆在目标图像中所在的平面区域，往往是不全面的。这是因为，在实际操作中，车辆具有一定的立体特征，而平面区域并不能很好的体现车辆的立体特征，例如车辆的摆放形态以及与图像拍摄设备的距离等。
举例来说，在车辆的车头、车尾或侧身均不是正对图像拍摄设备时，若通过车辆在目标图像中所在的平面区域来计算车辆与图像拍摄设备的距离，通常是不准确的，其中图像拍摄设备是对车辆进行拍摄得到目标图像的设备。参考图1所示为现有技术中通过目标图像获取特征的示意图，在目标图像中，其左下角的车辆与图像拍摄设备的距离较近的左侧车尾，如果根据车辆在图像中的区域中心来计算车辆与图像拍摄设备的距离，那么计算得到的距离应该是车辆内部与图像拍摄设备的距离，因此计算得到的距离要大于车辆与图像拍摄设备的实际距离，若根据计算得到的结果进行辅助驾驶，可能会导致较严重的后果。
基于此，在本发明实施例中，可以获取待处理图像，待处理图像中包括车辆，对待处理图像进行识别，得到车辆上的特征元素，其中特征元素可以是车辆的侧身元素和车端的元素，车端可以是车头或车尾，根据车辆的侧身元素和车端的元素在待处理图像中的位置，可以确定车辆的侧身区域和车端的区域，进而可以根据侧身区域获取侧身的区域特征，根据车端的区域获取车端的区域特征。因此，在本申请实施例中，先对待处理图像中的车辆进行了区域的划分，再在划分得到的区域中获取区域特征，相比于现有技术中只确定车辆在待处理图像中的位置，能够得到更全面的车辆特征。
下面结合附图，通过实施例来详细说明本发明实施例中一种车辆特征获取方法的具体实现方式。
参考图2所示为本申请实施例提供的一种车辆特征获取方法的流程图，该方法可以包括以下步骤。
S101，获取待处理图像。
在本申请实施例中，将目标图像作为待处理图像。其中，待处理图像是包括车辆的图像，可以通过图像拍摄设备进行拍摄，图像拍摄设备可以是照相机、摄像机、摄像头等。
在待处理图像中，可以仅包括车辆，也可以包括车辆之外的其他目标对象，待处理图像中可以包括一个车辆，也可以包括多个车辆。
S102，对待处理图像进行识别，得到车辆上的特征元素。
在本申请实施例中，车辆上的特征元素可以是侧身元素和车端的元素，其中车端可以是车头或车尾。车头的元素是可以辅助识别车头位置的元素，例如可以包括前车灯、前车窗、保险杠和前车牌等中的一项或多项，车尾的元素是可以辅助识别车尾位置的元素，例如可以包括后车灯、后车窗和后车牌等中的一项或多项，侧身元素是可以辅助识别侧身位置的元素，例如可以是车轮、侧车窗、后视镜和车门等中的一项或多项。
对待处理图像进行识别，得到车辆上的特征元素，可以具体为，对待处理图像进行分割，得到待处理图像中各个像素点的类别标签，其中，在待处理图像的分割图像中，同一特征元素的像素点的类别标签相同，根据类别标签得到车辆上的特征元素。对待处理图像进行分割可以通过预先训练得到的深度学习 神经网络进行，具体的，可以识别红色的椭圆形的区域为车灯区域，方形大面积的灰色区域为车窗区域，圆形具有白色射线状的区域为车轮等。
为了方便对分割后的图像进行处理，在对待处理图像进行分割后，可以为不同的类别标签预设不同的颜色，形成分割图像，例如车灯区域的类别标签对应的颜色可以是红色，车窗区域的类别标签对应的颜色可以是蓝色等。
在待处理图像包括车辆之外的其他目标对象时，为了提高图像分割的准确性，可以将待处理图像中对特征元素的识别分为两个步骤：对待处理图像进行识别，确定车辆所在的车辆区域，在待处理图像中截取车辆区域，形成目标图像，对目标图像进行识别，确定车辆区域中的特征元素。
具体的，可以通过预先训练得到的第一深度学习神经网络对待处理图像进行识别，得到第一分割图像，其中第一分割图像中各个像素点具有类别标签，且相同目标对象的像素点的类别标签相同，从而根据类别标签识别出待处理图像中的车辆区域；可以通过预先训练得到的第二深度学习神经网络对目标图像进行识别，得到第二分割图像，其中第二分割图像中各个像素点具有类别标签，且同一特征元素的类别标签相同，根据类别标签可以识别出车辆区域中的特征元素。
举例来说，参考图3所示为本申请实时提供的一种特征元素识别过程示意图，其中，图3(a)所示为示例性的待处理图像，通过第一深度学习神经网络对待处理进行分割，得到图3(b)所示的第一分割图像，其中，矩形框中的区域为车辆所在的车辆区域；对第一分割图像中的矩形框区域进行放大，得到参考图3(c)所示的车辆分割图像；根据车辆分割图像，对待处理图像进行截取，得到图3(d)所示的车辆图像；通过第二深度学习神经网络对车辆图像进行分割，得到图3(e)所示的第二分割图像，其中，两个圆形的区域为车辆上的车灯，一个矩形的区域为车辆上的车窗，当然，本申请实施例中，还可以识别出车胎或后视镜等其他元素。
S103，根据侧身元素和车端的元素在待处理图像中的位置，确定车辆的侧身区域和车端的区域。
在本申请实施例中，待处理图像中的车辆至少包括侧身元素和车端的元素，可以直接根据侧身元素和车端的元素确定车辆的侧身区域和车端的区域。
举例来说，参考图3(e)所示的第二分割图像，车尾的元素例如后车灯、 车窗等，侧身元素例如车轮、后视镜等，可以确定侧身元素中的车轮为后侧车轮，从而确定待处理图像中车尾的区域为包括车辆所在的矩形区域，侧身区域为车尾的区域的左边缘和后边缘。
在本申请实施例中，在确定车辆的侧身区域和车端的区域之前，也可以先确定车辆所在的车辆区域，以及侧身区域和车端的区域的分界线，再根据分界线，在车辆区域中确定侧身区域和车端的区域。
具体的，可以根据靠近侧身元素的车端的元素和靠近车端的元素的侧身元素在待处理图像中的位置，确定侧身区域和车端的区域的分界线。举例来说，图4所示为本申请实施例提供的一种区域确定过程示意图，其中，参考图4(a)所示的待处理图像，可以通过图像分割得到图4(b)所示的车辆区域(灰色矩形框)；对待处理图像进行图像分割，可以得到参考图4(c)所示的特征元素，其中，右侧两个椭圆的区域(黑色线条构成的椭圆区域)为车灯，右侧四边形的区域(黑色线条构成的四边形区域)为前车窗，左侧两个椭圆的区域(白色细线条构成的椭圆区域)为车轮，不规则形状的区域(白色细线条构成的不规则区域)为侧车窗。
通过分析可知，车端为车头，靠近车头的元素的侧身元素为待处理图像中的右车轮，以及待处理图像中侧车窗的右边缘；靠近侧身元素的车头的元素为前车窗的左边缘，以及车头的左车灯，因此，可以根据待处理图像中的右侧车轮、侧车窗的右侧边缘、前车窗的左侧边缘以及车头的左车灯确定车头的区域和侧身区域的分界线。
当然，车头的区域和侧身区域的分界线可以有多种确定方式，例如可以根据前车窗的左边缘和侧车窗的右边缘确定，也可以根据右车轮的右边缘和左车灯的左边缘确定。分界线可以是竖直的，也可以是倾斜的。在本申请实施例中，作为示例性的，可以通过右车轮的右边缘做竖直的直线，作为车头的区域和侧身区域的分界线，参考图4(d)所示的灰色线条。
在确定侧身区域和车端的区域的分界线后，可以根据确定出的分界线，以及侧身元素和车端的元素在待处理图像中的位置，在车辆区域中确定侧身区域和车端的区域。可以理解的是，侧身区域和车端的区域位于分界线两侧，侧身区域和车端的区域可以是矩形区域，也可以是平行四边形区域，还可以是其他可以表示区域的形状。
类比于分界线的确定方式，确定车辆区域中侧身区域和车端的区域，也可以有多种实现方式，举例来说，将通过右车轮的右边缘的竖直直线作为车头的区域和侧身区域的分界线，则参考图4(e)所示，可以以过车轮与地面的接触点的水平线作为车头的区域的下边缘，以分界线作为车头的区域的左边缘，以车顶所在水平线作为车头的区域的上边缘，以车辆区域的右边缘作为车头的区域的右边缘，四条边缘线条构成车头的区域(右侧白色粗线条构成的矩形框)；同理，以分界线作为侧身区域的右边缘，以车辆区域的左边缘作为侧身区域的左边缘，以车轮与地面的两个接触点连线的方向作为侧身区域的上边缘和下边缘的方向，其中，上边缘的右端点与车头的区域的左上端点(即车顶所在水平线和分界线的交点)重合，下边缘的右端点与车头的区域的左下端点(即车轮与地面的接触点)重合，四条边缘线条构成侧身区域(左侧白色粗线条构成的平行四边形框)。
为了便于数据存储，侧身区域和车端的区域可以通过六个顶点在待处理图像中的坐标来表示，由于分界线的确定方式可以有多种，侧身区域和车端的区域的确定方式也可以有多种，因此可以得到多组顶点坐标值，通过对多组顶点坐标值进行加权平均，得到加权顶点坐标值，表示侧身区域和车端的区域，以提高侧身区域和车端的区域的准确性。
S104，根据侧身区域获取侧身的区域特征，根据车端的区域获取车端的区域特征。
在本申请实施例中，侧身的区域特征可以是侧身区域的顶点、侧身区域的中心点或者侧身区域的边的中点等特征点在待处理图像中的位置，车端的区域特征包括车端的区域的顶点、车端的区域的中心点或者车端的区域的边的中点等特征点在待处理图像中的位置，例如车头的区域特征可以是车头的区域的顶点、车头的区域的中心点或者车头的区域的边的中点等特征点在待处理图像中的位置。可以理解的是，特征点在待处理图像中的位置，可以通过特征点在待处理图像中的坐标表示。
在获取到侧身的区域特征和车端的区域特征后，可以根据侧身区域的顶点和车端的区域特征的顶点，得到车辆的三维特征区域，车辆的三维特征区域可以包括车头的区域、车尾的区域、两边的侧身区域、车顶区域、车底区域。
在本申请实施例提供的一种车辆特征获取方法中，先获取待处理图像，待 处理图像中包括车辆，对待处理图像进行识别，得到车辆上的特征元素，其中特征元素可以是车辆的侧身元素和车端的元素，车端可以是车头或车尾，再根据车辆的侧身元素和车端的元素在待处理图像中的位置，可以确定车辆的侧身区域和车端的区域，进而可以根据侧身区域获取侧身的区域特征，根据车端的区域获取车端的区域特征。因此，先对待处理图像中的车辆进行了区域的划分，再在划分得到的区域中获取区域特征，相比于现有技术中只确定车辆在待处理图像中的位置，能够得到更全面的车辆特征。
在实际操作中，获取车辆特征是为了通过车辆特征进行相应的分析，以得到相关结果，作为示例性的，本申请实施例在获取车端的区域特征和侧身的区域特征后，还可以根据车端的区域特征和侧身的区域特征，以及待处理图像的拍摄参数，确定车辆与图像拍摄设备的相对位置。
其中，图像拍摄设备是对车辆进行拍摄得到待处理图像的设备，待处理图像的拍摄参数是指图像拍摄设备对车辆进行拍摄时的焦距、广角、位置、转角等参数。根据光学成像原理，可以根据待处理图像中特征点的位置以及待处理图像的拍摄参数，确定特征点与图像拍摄设备的相对位置，而待处理图像中车端的区域特征和侧身的区域特征可以全面的反映车辆的立体特征，因此，可以据此得到车辆与图像拍摄设备的相对位置。
作为一种可能的实现方式，根据车端的区域特征和侧身的区域特征，以及待处理图像的拍摄参数，确定车辆与图像拍摄设备的相对位置，可以具体为，将车端的区域和侧身区域中，朝向图像拍摄设备的区域确定为目标区域，根据目标区域的区域特征，以及待处理图像的拍摄参数，确定车辆与图像拍摄设备的相对位置。
具体的，可以根据车端的区域和侧身区域的面积和/或形状确定朝向图像拍摄设备的目标区域，例如将面积较大的区域确定为朝向图像拍摄设备的目标区域，或者将形状矩形的区域确定为朝向图像拍摄设备的目标区域。
参考图4(e)所示，车头的区域的面积大于侧身区域的面积，则可以认为车头的区域朝向图像拍摄设备，因此可以确定车头的区域为目标区域；而车头的区域的形状为矩形，而侧身区域的形状为平行四边形，因此可认为车头的区域朝向图像拍摄设备，因此可以确定车头的区域为目标区域。
目标区域的区域特征可以是目标区域的顶点、目标区域的中心点或者目标区域的边的中点等特征点在待处理图像中的位置，根据目标区域的区域特征以及待处理图像的拍摄参数，可以确定这些特征点与图像拍摄设备的相对位置，据此，可以得到车辆与图像拍摄设备的相对位置。
以图4(e)所示的车辆为例，目标区域为车头的区域，此时可以根据车头的区域的中心点在待处理图像中的位置，以及待处理图像的拍摄参数，确定车头的区域的中心点与图像拍摄设备的相对位置。
作为另一种可能的实现方式，根据车端的区域特征和侧身的区域特征，以及待处理图像的拍摄参数，确定车辆与图像拍摄设备的相对位置，可以具体为，根据车端的区域和侧身区域的分界线，以及待处理图像的拍摄参数，确定车辆与图像拍摄设备的相对位置。
具体的，可以确定分界线的中点，根据分界线的中点以及待处理图像的拍摄参数，确定车辆与图像拍摄设备的相对位置。这是因为车端的区域和侧身区域的分界线，往往是距离图像拍摄设备最近的，因此得到的车辆与图像拍摄设备的相对位置相对较准确。
基于本申请实施例提供的一种车辆特征获取方法，本申请实施例还提供了一种车辆特征获取装置，参考图5所示为本申请实施例提供的一种车辆特征获取装置的结构框图，该装置可以包括：
图像获取单元110，用于获取待处理图像，所述待处理图像中包括车辆；
图像识别单元120，用于对所述待处理图像进行识别，得到所述车辆上的特征元素；所述特征元素包括侧身元素和车端的元素，所述车端为车头或车尾；
区域确定单元130，用于根据所述侧身元素和所述车端的元素在所述待处理图像中的位置，确定所述车辆的侧身区域和车端的区域；
特征获取单元140，用于根据所述侧身区域获取所述侧身的区域特征，根据所述车端的区域获取所述车端的区域特征。
可选的，所述区域确定单元，包括：
车辆区域确定单元，用于在所述待处理图像中确定所述车辆所在的车辆区域；
分界线确定单元，用于根据靠近所述侧身元素的车端的元素和靠近所述车端的元素的侧身元素在所述待处理图像中的位置，确定侧身区域和车端的区域的分界线；
区域确定子单元，用于根据所述分界线，以及所述侧身元素和所述车端的元素在所述待处理图像中的位置，在所述车辆区域中确定侧身区域和车端的区域。
可选的，所述装置还包括：
相对位置确定单元，用于根据所述车端的区域特征和所述侧身的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与图像拍摄设备的相对位置。
可选的，所述相对位置确定单元，包括：
目标区域确定单元，用于将所述车端的区域和所述侧身区域中，朝向图像拍摄设备的区域确定为目标区域，所述图像拍摄设备用于对所述车辆进行拍摄得到待处理图像；
相对位置确定子单元，用于根据目标区域的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像拍摄设备的相对位置。
可选的，所述相对位置确定单元具体用于：
根据所述车端的区域和所述侧身区域的分界线，以及所述待处理图像的拍摄参数，确定所述车辆与图像拍摄设备的相对位置。
可选的，所述侧身的区域特征包括所述侧身区域的顶点、所述侧身区域的中心点或所述侧身区域的边的中点在所述待处理图像中的位置；所述车端的区域特征包括所述车端的区域的顶点、所述车端的区域的中心点或所述车端的区域的边的中点在所述待处理图像中的位置。
可选的，所述图像识别单元，包括：
车辆区域识别单元，用于在所述待处理图像中进行识别，确定所述车辆所在的车辆区域；
车辆图像获取单元，用于在所述待处理图像中截取所述车辆区域形成车辆图像；
特征元素识别单元，用于对所述车辆图像进行识别，确定所述车辆上的特 征元素。
可选的，所述车头的元素包括前车灯、前车窗、保险扛和前车牌中的一项或多项，所述车尾的元素包括后车灯、后车窗和后车牌中的一项或多项，所述侧身元素包括车轮、侧车窗、后视镜和车门中的一项或多项。
在本申请实施例提供的一种车辆特征获取装置中，先获取待处理图像，待处理图像中包括车辆，对待处理图像进行识别，得到车辆上的特征元素，其中特征元素可以是车辆的侧身元素和车端的元素，车端可以是车头或车尾，再根据车辆的侧身元素和车端的元素在待处理图像中的位置，可以确定车辆的侧身区域和车端的区域，进而可以根据侧身区域获取侧身的区域特征，根据车端的区域获取车端的区域特征。因此，先对待处理图像中的车辆进行了区域的划分，再在划分得到的区域中获取区域特征，相比于现有技术中只确定车辆在待处理图像中的位置，能够得到更全面的车辆特征。
本发明实施例中提到的“第一……”、“第一……”等名称中的“第一”只是用来做名字标识，并不代表顺序上的第一。该规则同样适用于“第二”等。
通过以上的实施方式的描述可知，本领域的技术人员可以清楚地了解到上述实施例方法中的全部或部分步骤可借助软件加通用硬件平台的方式来实现。基于这样的理解，本发明的技术方案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存储介质中，如只读存储器(英文：read-only memory，ROM)/RAM、磁碟、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，或者诸如路由器等网络通信设备)执行本发明各个实施例或者实施例的某些部分所述的方法。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于方法实施例和设备实施例而言，由于其基本相似于系统实施例，所以描述得比较简单，相关之处参见系统实施例的部分说明即可。以上所描述的设备及系统实施例仅仅是示意性的，其中作为分离部件说明的模块可以是或者也可以不是物理上分开的，作为模块显示的部件可以是或者也可以不是物理模块，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。本领域普 通技术人员在不付出创造性劳动的情况下，即可以理解并实施。
以上所述仅是本发明的优选实施方式，并非用于限定本发明的保护范围。应当指出，对于本技术领域的普通技术人员来说，在不脱离本发明的前提下，还可以作出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。

Claims (16)

  1. 一种车辆特征获取方法，其特征在于，所述方法包括：
    获取待处理图像，所述待处理图像中包括车辆；
    对所述待处理图像进行识别，得到所述车辆上的特征元素；所述特征元素包括侧身元素和车端的元素，所述车端为车头或车尾；
    根据所述侧身元素和所述车端的元素在所述待处理图像中的位置，确定所述车辆的侧身区域和车端的区域；
    根据所述侧身区域获取侧身的区域特征，根据所述车端的区域获取车端的区域特征。
  2. 根据权利要求1所述的方法，其特征在于，所述根据所述侧身元素和所述车端的元素在所述待处理图像中的位置，确定所述车辆的侧身区域和车端的区域，包括：
    在所述待处理图像中确定所述车辆所在的车辆区域；
    根据靠近所述侧身元素的车端的元素和靠近所述车端的元素的侧身元素在所述待处理图像中的位置，确定所述侧身区域和所述车端的区域的分界线；
    根据所述分界线，以及所述侧身元素和所述车端的元素在所述待处理图像中的位置，在所述车辆区域中确定所述侧身区域和所述车端的区域。
  3. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    根据所述车端的区域特征和所述侧身的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与图像拍摄设备的相对位置，所述图像拍摄设备用于对所述车辆进行拍摄得到所述待处理图像。
  4. 根据权利要求3所述的方法，其特征在于，所述根据所述车端的区域特征和所述侧身的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与图像拍摄设备的相对位置，包括：
    将所述车端的区域和所述侧身区域中，朝向所述图像拍摄设备的区域确定为目标区域；
    根据所述目标区域的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像拍摄设备的相对位置。
  5. 根据权利要求3所述的方法，其特征在于，所述根据所述车端的区域特征和所述侧身的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像获取设备的相对位置，包括：
    根据所述车端的区域和所述侧身区域的分界线，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像拍摄设备的相对位置。
  6. 根据权利要求1至5所述任意一项所述的方法，其特征在于，所述侧身的区域特征包括所述侧身区域的顶点、所述侧身区域的中心点或所述侧身区域的边的中点在所述待处理图像中的位置；所述车端的区域特征包括所述车端的区域的顶点、所述车端的区域的中心点或所述车端的区域的边的中点在所述待处理图像中的位置。
  7. 根据权利要求1至5所述任意一项所述的方法，其特征在于，所述对所述待处理图像进行识别，得到所述车辆上的特征元素，包括：
    在所述待处理图像中进行识别，确定所述车辆所在的车辆区域；
    在所述待处理图像中截取所述车辆区域形成车辆图像；
    对所述车辆图像进行识别，确定所述车辆上的特征元素。
  8. 根据权利要求1至5任意一项所述的方法，其特征在于，所述车头的元素包括前车灯、前车窗、保险扛和前车牌中的一项或多项，所述车尾的元素包括后车灯、后车窗和后车牌中的一项或多项，所述侧身元素包括车轮、侧车窗、后视镜和车门中的一项或多项。
  9. 一种车辆特征获取装置，其特征在于，所述装置包括：
    图像获取单元，用于获取待处理图像，所述待处理图像中包括车辆；
    图像识别单元，用于对所述待处理图像进行识别，得到所述车辆上的特征元素；所述特征元素包括侧身元素和车端的元素，所述车端为车头或车尾；
    区域确定单元，用于根据所述侧身元素和所述车端的元素在所述待处理图像中的位置，确定所述车辆的侧身区域和车端的区域；
    特征获取单元，用于根据所述侧身区域获取侧身的区域特征，根据所述车端的区域获取车端的区域特征。
  10. 根据权利要求9所述的装置，其特征在于，所述区域确定单元，包括：
    车辆区域确定单元，用于在所述待处理图像中确定所述车辆所在的车辆区域；
    分界线确定单元，用于根据靠近所述侧身元素的车端的元素和靠近所述车端的元素的侧身元素在所述待处理图像中的位置，确定所述侧身区域和所述车端的区域的分界线；
    区域确定子单元，用于根据所述分界线，以及所述侧身元素和所述车端的 元素在所述待处理图像中的位置，在所述车辆区域中确定所述侧身区域和所述车端的区域。
  11. 根据权利要求9所述的装置，其特征在于，所述装置还包括：
    相对位置确定单元，用于根据所述车端的区域特征和所述侧身的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与图像拍摄设备的相对位置，所述图像拍摄设备用于对所述车辆进行拍摄得到所述待处理图像。
  12. 根据权利要求11所述的装置，其特征在于，所述相对位置确定单元，包括：
    目标区域确定单元，用于将所述车端的区域和所述侧身区域中，朝向所述图像拍摄设备的区域确定为目标区域；
    相对位置确定子单元，用于根据所述目标区域的区域特征，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像拍摄设备的相对位置。
  13. 根据权利要求11所述的装置，其特征在于，所述相对位置确定单元具体用于：
    根据所述车端的区域和所述侧身区域的分界线，以及所述待处理图像的拍摄参数，确定所述车辆与所述图像拍摄设备的相对位置。
  14. 根据权利要求9至13所述任意一项所述的装置，其特征在于，所述侧身的区域特征包括所述侧身区域的顶点、所述侧身区域的中心点或所述侧身区域的边的中点在所述待处理图像中的位置；所述车端的区域特征包括所述车端的区域的顶点、所述车端的区域的中心点或所述车端的区域的边的中点在所述待处理图像中的位置。
  15. 根据权利要求9至13所述任意一项所述的装置，其特征在于，所述图像识别单元，包括：
    车辆区域识别单元，用于在所述待处理图像中进行识别，确定所述车辆所在的车辆区域；
    车辆图像获取单元，用于在所述待处理图像中截取所述车辆区域形成车辆图像；
    特征元素识别单元，用于对所述车辆图像进行识别，确定所述车辆上的特征元素。
  16. 根据权利要求9至13任意一项所述的装置，其特征在于，所述车头的元素包括前车灯、前车窗、保险扛和前车牌中的一项或多项，所述车尾的元 素包括后车灯、后车窗和后车牌中的一项或多项，所述侧身元素包括车轮、侧车窗、后视镜和车门中的一项或多项。
PCT/CN2019/084537 2018-09-27 2019-04-26 一种车辆特征获取方法及装置 Ceased WO2020062856A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
US17/280,191 US12002269B2 (en) 2018-09-27 2019-04-26 Vehicle feature acquisition method and device
EP19864666.3A EP3859593B1 (en) 2018-09-27 2019-04-26 Vehicle feature acquisition method and device
JP2021542238A JP7190583B2 (ja) 2018-09-27 2019-04-26 車両特徴の取得方法及び装置

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811132325.X 2018-09-27
CN201811132325.XA CN109389064B (zh) 2018-09-27 2018-09-27 一种车辆特征获取方法及装置

Publications (1)

Publication Number Publication Date
WO2020062856A1 true WO2020062856A1 (zh) 2020-04-02

Family

ID=65418211

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/084537 Ceased WO2020062856A1 (zh) 2018-09-27 2019-04-26 一种车辆特征获取方法及装置

Country Status (5)

Country Link
US (1) US12002269B2 (zh)
EP (1) EP3859593B1 (zh)
JP (1) JP7190583B2 (zh)
CN (1) CN109389064B (zh)
WO (1) WO2020062856A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114194184A (zh) * 2020-09-17 2022-03-18 现代自动车株式会社 车辆和控制车辆的方法

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109389064B (zh) * 2018-09-27 2021-02-23 东软睿驰汽车技术（沈阳）有限公司 一种车辆特征获取方法及装置
CN110059566A (zh) * 2019-03-20 2019-07-26 东软睿驰汽车技术（沈阳）有限公司 一种图像识别方法及装置
CN110084230B (zh) * 2019-04-11 2021-05-28 北京百度网讯科技有限公司 基于图像的车身方向检测方法和装置
CN110765929A (zh) * 2019-10-21 2020-02-07 东软睿驰汽车技术（沈阳）有限公司 一种车辆障碍物检测方法及装置
CN110765961A (zh) * 2019-10-29 2020-02-07 上海眼控科技股份有限公司 车辆制动状态判断方法、装置、计算机设备和存储介质
CN110796078A (zh) * 2019-10-29 2020-02-14 上海眼控科技股份有限公司 车辆的灯光检测方法、装置、电子设备及可读存储介质
CN111556362A (zh) * 2020-03-19 2020-08-18 上海万面智能科技有限公司 一种车身广告植入方法、装置、电子设备及存储介质
CN111461026B (zh) * 2020-04-02 2024-03-12 北京爱笔科技有限公司 车辆姿态识别方法、停车场内车辆状态的检测方法及装置
CN111461027A (zh) * 2020-04-02 2020-07-28 北京爱笔科技有限公司 一种车辆检测方法及装置、车牌识别方法及装置
CN111476169B (zh) * 2020-04-08 2023-11-07 智慧互通科技股份有限公司 一种基于视频帧的复杂场景路侧停车行为识别方法
CN113436257B (zh) * 2021-06-09 2023-02-10 同济大学 一种基于道路几何信息的车辆位置实时检测方法
CN114323583B (zh) * 2021-12-21 2024-06-04 广汽本田汽车有限公司 一种车辆灯光检测方法、装置、设备及系统
CN115239976A (zh) * 2022-06-30 2022-10-25 北京万集科技股份有限公司 车辆信息检测方法和车辆信息检测系统

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104021379A (zh) * 2014-06-09 2014-09-03 北京航空航天大学 一种车辆辅助驾驶中车辆间距的计算方法
US20160375904A1 (en) * 2015-06-29 2016-12-29 The Regents Of The University Of California Multiple-parts based vehicle detection integrated with lane detection for improved computational efficiency and robustness
CN107256633A (zh) * 2017-05-09 2017-10-17 西安理工大学 一种基于单目摄像头三维估计的车型分类方法
KR20180061695A (ko) * 2016-11-30 2018-06-08 주식회사 만도 휠 검출을 통한 차량 옆면 인식 방법 및 장치
CN108241822A (zh) * 2016-12-23 2018-07-03 杭州海康威视数字技术股份有限公司 一种车辆类型的识别方法及装置
CN108388871A (zh) * 2018-02-28 2018-08-10 中国计量大学 一种基于车身回归的车辆检测方法
CN109389064A (zh) * 2018-09-27 2019-02-26 东软睿驰汽车技术（沈阳）有限公司 一种车辆特征获取方法及装置

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106023220B (zh) * 2016-05-26 2018-10-19 史方 一种基于深度学习的车辆外观部件图像分割方法
JP6678552B2 (ja) 2016-09-30 2020-04-08 株式会社東芝 車種判別装置および車種判別方法
CN108170708B (zh) * 2017-11-23 2021-03-30 杭州大搜车汽车服务有限公司 一种车辆实体识别方法、电子设备、存储介质、系统
CN108364010B (zh) * 2018-03-08 2022-03-22 广东工业大学 一种车牌识别方法、装置、设备及计算机可读存储介质
CN108491782B (zh) 2018-03-16 2020-09-08 重庆大学 一种基于行车图像采集的车辆识别方法
CN108509907B (zh) 2018-03-30 2022-03-15 北京市商汤科技开发有限公司 车灯检测方法、实现智能驾驶的方法、装置、介质及设备
US10176405B1 (en) * 2018-06-18 2019-01-08 Inception Institute Of Artificial Intelligence Vehicle re-identification techniques using neural networks for image analysis, viewpoint-aware pattern recognition, and generation of multi- view vehicle representations
US20200020117A1 (en) * 2018-07-16 2020-01-16 Ford Global Technologies, Llc Pose estimation

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104021379A (zh) * 2014-06-09 2014-09-03 北京航空航天大学 一种车辆辅助驾驶中车辆间距的计算方法
US20160375904A1 (en) * 2015-06-29 2016-12-29 The Regents Of The University Of California Multiple-parts based vehicle detection integrated with lane detection for improved computational efficiency and robustness
KR20180061695A (ko) * 2016-11-30 2018-06-08 주식회사 만도 휠 검출을 통한 차량 옆면 인식 방법 및 장치
CN108241822A (zh) * 2016-12-23 2018-07-03 杭州海康威视数字技术股份有限公司 一种车辆类型的识别方法及装置
CN107256633A (zh) * 2017-05-09 2017-10-17 西安理工大学 一种基于单目摄像头三维估计的车型分类方法
CN108388871A (zh) * 2018-02-28 2018-08-10 中国计量大学 一种基于车身回归的车辆检测方法
CN109389064A (zh) * 2018-09-27 2019-02-26 东软睿驰汽车技术（沈阳）有限公司 一种车辆特征获取方法及装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3859593A4 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114194184A (zh) * 2020-09-17 2022-03-18 现代自动车株式会社 车辆和控制车辆的方法

Also Published As

Publication number Publication date
EP3859593B1 (en) 2025-02-26
EP3859593A4 (en) 2022-06-29
US20220004784A1 (en) 2022-01-06
US12002269B2 (en) 2024-06-04
JP2022501756A (ja) 2022-01-06
CN109389064A (zh) 2019-02-26
JP7190583B2 (ja) 2022-12-15
CN109389064B (zh) 2021-02-23
EP3859593A1 (en) 2021-08-04

Similar Documents

Publication Publication Date Title
WO2020062856A1 (zh) 一种车辆特征获取方法及装置
CN107577988B (zh) 实现侧方车辆定位的方法、装置及存储介质、程序产品
EP2237988B1 (en) Object detection and recognition system
CN104778444B (zh) 道路场景下车辆图像的表观特征分析方法
WO2020124988A1 (zh) 一种基于视觉的车位检测方法及装置
CN104899554A (zh) 一种基于单目视觉的车辆测距方法
WO2020248910A1 (zh) 目标检测方法和装置
US8098933B2 (en) Method and apparatus for partitioning an object from an image
JP2002352225A (ja) 障害物検出装置及びその方法
US11164012B2 (en) Advanced driver assistance system and method
US11900696B2 (en) Method and system for processing a plurality of images so as to detect lanes on a road
US11236991B2 (en) Method for determining a current distance and/or a current speed of a target object based on a reference point in a camera image, camera system and motor vehicle
CN105059190B (zh) 基于视觉的汽车开门碰撞预警装置及方法
US20240103525A1 (en) Vehicle and control method thereof
Chang et al. An efficient method for lane-mark extraction in complex conditions
JP4344860B2 (ja) ステレオ画像を用いた道路平面領域並びに障害物検出方法
CN110969135B (zh) 自然场景中的车标识别方法
CN117152210B (zh) 基于动态观测视场角的图像动态追踪方法及相关装置
WO2019052320A1 (zh) 监控方法、装置、系统、电子设备及计算机可读存储介质
WO2020187311A1 (zh) 一种图像识别方法及装置
CN107992788A (zh) 识别交通灯的方法、装置及车辆
CN108399357B (zh) 一种人脸定位方法及装置
Ponsa et al. On-board image-based vehicle detection and tracking
CN115995142A (zh) 基于可穿戴装置的驾驶训练提醒方法及可穿戴装置
CN107992789A (zh) 识别交通灯的方法、装置及车辆

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: 19864666

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021542238

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: 2019864666

Country of ref document: EP

Effective date: 20210428