Anti-collision vehicle recognition camera system and method

Abstract

The present application relates to an anti-collision vehicle recognition camera system and method, which realizes the dynamic information extraction of the image by receiving the continuous images taken by the camera. Since the spacing between the mounting screws of the car license plate can be used as a distance standard reference, and the spectral information of the mounting screws of the car license plate is also referenced, the division of the license plate position in the image can be determined based on the spectral layer of the corresponding spectral information in the image. Depicting the area where the license plate is located in the original image is conducive to obtaining the image corresponding to each license plate area. When the license plate information in the image is determined, the license plate information of the adjacent one can be found in the subsequent image. Based on the determined license plate information, the position relationship and relative speed relationship between the vehicles corresponding to these license plates and the vehicle equipped with the camera can be judged through the vector velocity model, and then the vehicle anti-collision warning can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of recognition algorithms, and in particular to an anti-collision vehicle recognition camera system and method. Background Art

[0002] As the degree of automobile automation increases, more and more cars are equipped with cameras. Compared with traditional millimeter-wave radars, cameras have richer picture information. In addition, the image information extraction algorithm based on cameras is relatively mature. For this reason, more and more automobile manufacturers have shifted the focus of research and development of automobile safety protection warning technology to image information extraction from cameras. However, traditional image information extraction technology is difficult to adapt to the needs of dynamic information extraction. The simplification of dynamic information extraction of images and the rapid application of extracted information are the industry difficulties in using image information for automobile safety protection warning. For this reason, it is necessary to propose an anti-collision vehicle recognition camera system and method in view of the defect that traditional image information extraction technology is difficult to apply to automobile safety protection warning. Summary of the invention

[0003] Based on this, it is necessary to propose an anti-collision vehicle recognition camera system and method to address the defect that traditional image information extraction technology is difficult to apply to automobile safety protection warning.

[0004] The present application provides an anti-collision vehicle recognition camera method, comprising:

[0005] Receive images captured by the camera;

[0006] Based on the spectral information of the mounting screws of the automobile license plate, extracting the spectral layer of the corresponding spectral information in the image;

[0007] Using the highlight part of the spectral layer, the image area of ​​the license plate position is divided to form the license plate area;

[0008] Describe the area where the license plate is located in the original image;

[0009] Inputting at least one image of the area where the license plate is located into a shape recognition algorithm to obtain information corresponding to each image of the area where the license plate is located;

[0010] Select the image of the area with the largest license plate;

[0011] Based on the vector velocity model, determine the vector displacement of the image of the area where the license plate is located;

[0012] If the vector displacement of the image of the area where the license plate is located points to the camera, it is further determined whether the rate of the vector displacement is greater than or equal to the rate threshold;

[0013] If the velocity of the vector displacement is greater than or equal to the velocity threshold, a collision warning is issued;

[0014] If the speed of the vector displacement is less than the speed threshold, returning to the vector speed model to determine the vector displacement of the image of the area where the license plate is located;

[0015] If the vector displacement of the image of the area where the license plate is located does not point to the camera, the image of the area where the license plate is located is included in the security folder, and the image of the area where the license plate is located with the largest selected area is returned.

[0016] The present application also provides an anti-collision vehicle recognition camera system, comprising:

[0017] A processor, used to execute the anti-collision vehicle recognition and camera method;

[0018] A camera is communicatively connected to the processor.

[0019] The present application relates to an anti-collision vehicle recognition camera system and method, which realizes the dynamic information extraction of the image by receiving the continuous images taken by the camera. Since the spacing between the mounting screws of the car license plate can be used as a distance standard reference, and the spectral information of the mounting screws of the car license plate is also referenced, the division of the license plate position in the image can be determined based on the spectral layer of the corresponding spectral information in the image. Depicting the area where the license plate is located in the original image is conducive to obtaining the image corresponding to each license plate area. When the license plate information in the image is determined, the license plate information of the adjacent one can be found in the subsequent image. Based on the determined license plate information, the position relationship and relative speed relationship between the vehicles corresponding to these license plates and the vehicle equipped with the camera can be judged through the vector velocity model, and then the vehicle anti-collision warning can be realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings constituting a part of this application are used to provide a further understanding of this application, so that other features, purposes and advantages of this application become more obvious. The illustrative embodiment drawings and their descriptions of this application are used to explain this application and do not constitute an improper limitation on this application.

[0021] Figure 1 A method flow chart of an anti-collision vehicle recognition camera method provided in one embodiment of the present application.

[0022] Figure 2 A license plate position diagram of an anti-collision vehicle recognition camera method provided in one embodiment of the present application.

[0023] Figure 3 A diagram showing the relationship between image size and position of an anti-collision vehicle recognition camera method provided in one embodiment of the present application.

[0024] Figure 4 A device connection diagram of an anti-collision vehicle recognition camera system provided in one embodiment of the present application.

[0025] Reference numerals:

[0026] 100-processor; 200-camera. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] The present application provides an anti-collision vehicle recognition and imaging method.

[0029] like Figure 1 As shown, in one embodiment of the present application, a method for anti-collision vehicle recognition and camera shooting includes:

[0030] S100, receiving an image captured by a camera.

[0031] S200 , based on the spectral information of the mounting screws of the automobile license plate, extracting a spectral layer of the corresponding spectral information in the image.

[0032] S300, using the highlight part of the spectral layer, the image region of the license plate position is divided to form a license plate area.

[0033] S400, depicting the area where the license plate is located in the original image.

[0034] S500: Inputting at least one image of the area where the license plate is located into a shape recognition algorithm to obtain information corresponding to each image of the area where the license plate is located.

[0035] Specifically, shape recognition is to recognize numbers by analyzing the shape features in an image. Commonly used algorithms include Hough transform and contour analysis. These algorithms extract contour information by corroding and dilating the binary image, and perform shape description and matching.

[0036] S600, select the image of the area where the license plate with the largest area is located.

[0037] S700, determining the vector displacement of the image of the area where the license plate is located based on the vector velocity model.

[0038] S810: If the vector displacement of the image of the area where the license plate is located points to the camera, further determine whether the rate of the vector displacement is greater than or equal to a rate threshold.

[0039] S820: If the speed of the vector displacement is greater than or equal to the speed threshold, a collision warning is issued.

[0040] S830: If the speed of the vector displacement is less than the speed threshold, return to the vector speed model to determine the vector displacement of the image of the area where the license plate is located.

[0041] S840, if the vector displacement of the image of the area where the license plate is located does not point to the camera, the image of the area where the license plate is located is included in the security folder, and the image of the area where the license plate is located with the largest selected area is returned.

[0042] The present application relates to an anti-collision vehicle recognition camera method, which realizes the dynamic information extraction of the image by receiving the continuous images taken by the camera. Since the spacing between the mounting screws of the automobile license plate can be used as a distance standard reference, and the spectral information of the mounting screws of the automobile license plate is also referenced, the division of the license plate position in the image can be determined based on the spectral layer of the corresponding spectral information in the image. Depicting the area where the license plate is located at the original image is conducive to obtaining the image corresponding to each license plate area. When the license plate information in the image is determined, the license plate information of the adjacent one can be found in the subsequent image. Based on the determined license plate information, the position relationship and relative speed relationship between the vehicles corresponding to these license plates and the vehicle equipped with the camera can be judged through the vector velocity model, and then the vehicle anti-collision warning can be realized.

[0043] In one embodiment of the present application, before S600, the following steps are included:

[0044] S611, select an image of the area where the license plate is located.

[0045] S612, counting the number of pixels occupied by the image of the area where the license plate is located.

[0046] S613, returning to the process of selecting an image of an area where a license plate is located, until all images of areas where license plates are located in the image are selected.

[0047] S614, arranging all images of the area where the license plate is located in order of the number of pixels from small to large.

[0048] Specifically, generally speaking, under the same viewing angle and with the same volume of the actual object, the farther the actual object is, the fewer pixels it occupies in the same photo. In this embodiment, the license plate with fewer pixels is farther away from the camera. The risk of collision between all vehicles with license plates farther away and the vehicle where the camera is located is smaller.

[0049] In one embodiment of the present application, before S600, the method further includes:

[0050] S621, select an image of the area where the license plate is located.

[0051] S622, orthogonalize the image of the area where the license plate is located to obtain an orthogonalized image.

[0052] S623, based on the conversion relationship between the distance and the image size, determine the actual distance between the orthogonalized image and the camera.

[0053] S624, returning to the process of selecting an image of the area where a license plate is located, until all images of the areas where license plates are located are selected.

[0054] Specifically, Figure 2 and Figure 3 As shown, the image of the area where the license plate is located is mostly a parallelogram or a rectangle. When the image of the area where the license plate is located is a rectangle, the image of the area where the license plate is located is an orthogonalized image.

[0055] When the image of the license plate area is a parallelogram:

[0056] In one embodiment of the present application, S622 includes:

[0057] S631, selecting an edge intersection point of the image of the area where the license plate is located.

[0058] S632: Connect the edge intersection point with the geometric center of the image captured by the camera.

[0059] S633, returning an edge intersection point of the image of the area where the license plate is located, until all edge intersection points are selected.

[0060] S634, based on the width and length between the mounting screws of the automobile license plate, determine the relative parameters between the geometric center of the image captured by the camera and the geometric center of the actual position of the area where the license plate is located.

[0061] S635, analyzing relative parameters, width and length between mounting screws of the car license plate, and obtaining an orthogonalized image.

[0062] Specifically, if the image of the area where the license plate is located has less than three edge intersections, it means that the license plate is blocked by an obstacle. In this case, the vehicle with a license plate with less than three edge intersections has a lower collision risk with the camera-mounted vehicle.

[0063] More specifically, when the image of the area where the license plate is located has three or four edge intersections, the line connecting the geometric center of the image taken by the camera and the edge intersection is the projection of the spatial line connecting the edge intersections of the camera entity and the license plate entity on the license plate plane.

[0064] In this embodiment, the line connecting the geometric center of the image taken by the camera and the edge intersection can approximately obtain the relative parameters of the geometric center of the image taken by the camera and the actual position of the license plate area when there are three edge intersections.

[0065] The relevant parameters are the approximate polar angle and the approximate azimuth angle.

[0066] In another embodiment, when there are four edge intersections in the image of the area where the license plate is located, the angle between the geometric center of the parallelogram formed by the four edge intersections and the geometric center of the image taken by the camera and the length direction of the license plate in the image is the azimuth angle, and the angle between the connecting line and the width direction of the license plate in the image is the polar angle.

[0067] The determined azimuth and polar angles can be used to orthogonalize the license plate in the image to form a rectangle orthogonal to the camera. Finally, the width and length between the mounting screws of the car license plate can be used to determine the actual distance between the license plate and the camera.

[0068] In one embodiment of the present application, S700 includes:

[0069] S711, select a system time.

[0070] S712, receiving the image of the area where the license plate is located under the system time.

[0071] S713, based on a preset time length, receiving an image of the area where the license plate is located a preset time length after the system time.

[0072] S714, if the pixels occupied by the image of the area where the license plate is located increase, determine that the vector displacement of the image of the area where the license plate is located points to the camera.

[0073] S715, if the pixels occupied by the image of the area where the license plate is located remain unchanged, determine that the vector displacement of the image of the area where the license plate is located is a zero vector; define the zero vector as a relatively static vector displacement vector.

[0074] S716, if the pixels occupied by the image of the area where the license plate is located decrease, it is determined that the vector displacement of the image of the area where the license plate is located does not point to the camera.

[0075] S721, calling the actual distance of the image of the area where the license plate is located in the system time; defining the actual distance as the first distance.

[0076] S722, calling the actual distance of the image of the area where the license plate is located after the system time for a preset time length; defining the actual distance as the second distance.

[0077] S723: Obtain a distance difference of the second distance minus the first distance.

[0078] S724, derive the distance difference from the preset time length to obtain the speed of the license plate.

[0079] Specifically, in a simplified collision avoidance warning system, when a vehicle corresponding to a license plate travels away from a camera, the vehicle equipped with the camera is safe.

[0080] Regardless of whether they are traveling in the same direction, the relative distance between the front vehicle and the vehicle where the camera is located is getting larger and larger, that is, the vector of the relative speed between the front vehicle and the vehicle where the camera is located is away from the vehicle where the camera is located.

[0081] When the pixels occupied by the image of the area where the license plate is located increase, it means that the vehicle corresponding to the license plate is approaching the vehicle where the camera is located. In this case, there may be a situation where two vehicles meet. During the meeting process, according to this embodiment, the relative speed of the two vehicles will infinitely tend to zero, and the two vehicles will have a minimum parallel distance.

[0082] When the relative speed exceeds the threshold, a collision safety warning needs to be issued.

[0083] In one embodiment of the present application, S700 includes:

[0084] S731, establish a spherical coordinate system about the camera.

[0085] S732, receiving an image of the area where the license plate is located at a moment.

[0086] S733, a first line is drawn to connect the midpoint in the length direction of the image of the area where the license plate is located with the geometric center of the image taken by the camera.

[0087] S734, a second line is drawn between the midpoint in the width direction of the image of the area where the license plate is located and the geometric center of the image taken by the camera.

[0088] S735, defining the angle between the length direction of the image of the area where the license plate is located and the extension direction of the first connecting line as a direction angle.

[0089] S736, defining the angle between the width direction of the image of the area where the license plate is located and the extension direction of the second connecting line as a polar angle.

[0090] S737, calling the actual distance of the image of the area where the license plate is located at that moment.

[0091] S738, generating the spherical coordinates of the image of the area where the license plate is located at that moment.

[0092] S741, define an interval time.

[0093] S742, receiving an image of the area where the license plate is located at an interval after the moment.

[0094] S743, based on the image of the area where the license plate is located, generate the spherical coordinates of the image of the area where the license plate is located at an interval time after the moment.

[0095] S751, determining the vector speed of the license plate based on the interval time, the spherical coordinates at the moment, and the spherical coordinates at an interval time after the moment.

[0096] Specifically, in this embodiment, the camera is used as the origin of the spherical coordinates, and the different spatial positions of the same license plate relative to the origin at different times are used to solve the vector displacement between the camera and the license plate.

[0097] Based on the known interval time and different spatial positions of the same license plate, the relative vector displacement of the license plate can be obtained. The direction and rate of the vector displacement can help the processor determine whether to issue a collision warning.

[0098] The present application provides an anti-collision vehicle recognition camera system.

[0099] like Figure 4 As shown, in one embodiment of the present application, an anti-collision vehicle recognition camera system includes a processor 100 and a camera 200 .

[0100] The processor 100 is used to execute the anti-collision vehicle recognition and imaging method.

[0101] The camera 200 is in communication connection with the processor 100 .

[0102] The present application relates to an anti-collision vehicle recognition camera system, in which the processor 100 realizes the dynamic information extraction of the image by receiving the continuous images taken by the camera 200. Since the spacing between the mounting screws of the automobile license plate can be used as a distance standard reference, and the spectral information of the mounting screws of the automobile license plate is also of reference, the division of the license plate position in the image can be determined based on the spectral layer of the corresponding spectral information in the image. Depicting the area where the license plate is located at the original image is conducive to obtaining the image corresponding to each license plate area. When the license plate information in the image is determined, the license plate information of the adjacent one can be found in the subsequent images taken. Based on the determined license plate information, the position relationship and relative speed relationship between the vehicles corresponding to these license plates and the vehicles equipped with the camera can be judged through the vector velocity model, so as to realize the vehicle anti-collision warning.

[0103] The technical features of the above-described embodiments may be arbitrarily combined, and the execution order of the method steps is not limited. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0104] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for anti-collision vehicle recognition and camera shooting, characterized in that: include: Receive images captured by the camera; Based on the spectral information of the mounting screws of the automobile license plate, extracting the spectral layer of the corresponding spectral information in the image; Using the highlight part of the spectral layer, the image area of ​​the license plate position is divided to form the license plate area; Describe the area where the license plate is located in the original image; Inputting at least one image of the area where the license plate is located into a shape recognition algorithm to obtain information corresponding to each image of the area where the license plate is located; Select the image of the area with the largest license plate; Based on the vector velocity model, determine the vector displacement of the image of the area where the license plate is located; If the vector displacement of the image of the area where the license plate is located points to the camera, it is further determined whether the rate of the vector displacement is greater than or equal to the rate threshold; If the velocity of the vector displacement is greater than or equal to the velocity threshold, a collision warning is issued; If the speed of the vector displacement is less than the speed threshold, returning to the vector speed model to determine the vector displacement of the image of the area where the license plate is located; If the vector displacement of the image of the area where the license plate is located does not point to the camera, the image of the area where the license plate is located is included in the security folder, and the image of the area where the license plate is located with the largest selected area is returned.

2. The anti-collision vehicle recognition camera method according to claim 1, characterized in that: Before selecting the image of the area where the license plate with the largest area is located, the method includes: Select an image of the area where the license plate is located; Count the number of pixels occupied by the image of the area where the license plate is located; Returning to the step of selecting an image of an area where a license plate is located, until all images of areas where license plates are located in the image are selected; Arrange all the images of the license plate area in order from small to large according to the number of pixels.

3. The anti-collision vehicle recognition camera method according to claim 2, characterized in that: Before selecting the image of the area where the license plate with the largest area is located, the method further includes: Select an image of the area where the license plate is located; Orthogonalize the image of the area where the license plate is located to obtain an orthogonalized image; Based on the conversion relationship between distance and image size, determining the actual distance between the orthogonalized image and the camera; Return to the process of selecting an image of the area where a license plate is located until all images of the areas where license plates are located are selected.

4. The anti-collision vehicle recognition camera method according to claim 3, characterized in that: The step of orthogonalizing the image of the area where the license plate is located to obtain an orthogonalized image includes: Select an edge intersection point of the image where the license plate is located; Connect the edge intersection point with the geometric center of the image captured by the camera; Return an edge intersection point of the image of the area where the selected license plate is located until all edge intersection points are selected; Based on the width and length between the mounting screws of the automobile license plate, relative parameters between the geometric center of the image captured by the camera and the geometric center of the actual position of the area where the license plate is located are determined; Analyze the relative parameters, width and length between the mounting screws of the car license plate, and obtain the orthogonalized image.

5. The anti-collision vehicle recognition camera method according to claim 4, characterized in that: The step of determining the vector displacement of the image of the area where the license plate is located based on the vector velocity model includes: Select a system time; Receive an image of the area where the license plate is located at the system time; Based on a preset time length, receiving an image of the area where the license plate is located for a preset time length after the system time; If the pixels occupied by the image of the area where the license plate is located increase, it is determined that the vector displacement of the image of the area where the license plate is located points to the camera; If the pixels occupied by the image of the area where the license plate is located remain unchanged, the vector displacement of the image of the area where the license plate is located is determined to be a zero vector; the zero vector is defined as a relatively static vector displacement vector; If the pixels occupied by the image of the area where the license plate is located decrease, it is determined that the vector displacement of the image of the area where the license plate is located does not point to the camera.

6. The anti-collision vehicle recognition camera method according to claim 5, characterized in that: The step of determining the vector displacement of the image of the area where the license plate is located based on the vector velocity model also includes: Call the actual distance of the image of the area where the license plate is located in the system time; define the actual distance as the first distance; Call the actual distance of the image of the area where the license plate is located after the system time for a preset time length; define the actual distance as the second distance; Obtaining a distance difference of the second distance minus the first distance; The speed of the license plate is obtained by dividing the distance difference by the preset time length.

7. The anti-collision vehicle recognition camera method according to claim 4, characterized in that: The step of determining the vector displacement of the image of the area where the license plate is located based on the vector velocity model includes: Establish a spherical coordinate system about the camera; Receive an image of the area where the license plate is located at a moment; A first line connecting the midpoint in the length direction of the image of the area where the license plate is located and the geometric center of the image taken by the camera; A second line connecting the midpoint in the width direction of the image of the area where the license plate is located and the geometric center of the image captured by the camera; The included angle between the length direction of the image of the area where the license plate is located and the extension direction of the first connecting line is defined as the direction angle; The angle between the width direction of the image of the area where the license plate is located and the extension direction of the second connecting line is defined as the polar angle; The actual distance of the image of the area where the license plate is located at that moment; Generate the spherical coordinates of the image of the area where the license plate is located at that moment.

8. The anti-collision vehicle recognition camera method according to claim 7, characterized in that: The step of determining the vector displacement of the image of the area where the license plate is located based on the vector velocity model also includes: Define an interval time; receiving an image of the area where the license plate is located at an interval after the moment; Based on the image of the area where the license plate is located, the spherical coordinates of the image of the area where the license plate is located at an interval time after the moment are generated.

9. The anti-collision vehicle recognition camera method according to claim 8, characterized in that: The step of determining the vector displacement of the image of the area where the license plate is located based on the vector velocity model also includes: The vector velocity of the license plate is determined based on the interval time, the spherical coordinates at the moment, and the spherical coordinates at an interval time after the moment.

10. An anti-collision vehicle recognition camera system, characterized in that: include: A processor, configured to execute the anti-collision vehicle recognition and imaging method according to any one of claims 1 to 9; A camera is communicatively connected to the processor.

Images