Radar result detection method, device, equipment and medium

Abstract

The invention discloses a radar result detection method and device, equipment and a medium. The method comprises the following steps: acquiring an obstacle point cloud detection result in a to-be-detected area of a to-be-detected vehicle sent by a radar; performing target detection on the obtained image of the recognizable area to obtain an obstacle image detection result; according to the obstacle image detection result, determining an obstacle space detection result; and checking the obstacle point cloud detection result according to the obstacle space detection result. According to the embodiment of the invention, the accuracy and efficiency of radar result detection can be improved.

Description

Technical Field

[0001] The present invention relates to the field of data detection technology, and in particular to a radar result detection method, device, equipment and medium. Background Art

[0002] With the rapid development of science and technology, various types of vehicles appear in all aspects of people's lives. During the driving process of the vehicle, various types of obstacles will appear around the vehicle, affecting the driving of the vehicle. Radar can now be used to sense obstacles around the vehicle to warn the driver. In order to make the obstacle point cloud sensed by radar more accurate, various parameters in the obstacle detection algorithm need to be continuously adjusted.

[0003] Currently, various parameters in the obstacle detection algorithm are adjusted based on multiple human experiments and experimental results.

[0004] However, traffic conditions are complex, the number of experiments is large, the amount of calculation is large, and the efficiency of human evaluation of experimental results is low. Summary of the invention

[0005] The present invention provides a radar result detection method, device, equipment and medium to improve the accuracy of radar result detection.

[0006] In a first aspect, an embodiment of the present invention provides a radar result detection method, the method comprising:

[0007] Obtaining the obstacle point cloud detection result in the detection area of ​​the vehicle to be detected sent by the radar;

[0008] Performing target detection on the acquired image of the identifiable area to obtain obstacle image detection results;

[0009] Determine obstacle space detection results based on obstacle image detection results;

[0010] The obstacle point cloud detection results are verified according to the obstacle space detection results.

[0011] In a second aspect, an embodiment of the present invention further provides a radar result detection device, the device comprising:

[0012] A point cloud data acquisition module is used to obtain the obstacle point cloud detection results in the detection area of ​​the vehicle to be detected sent by the radar;

[0013] An image data acquisition module is used to perform target detection on the acquired image of the identifiable area to obtain an obstacle image detection result;

[0014] The spatial data acquisition module is used to determine the obstacle space detection result according to the obstacle image detection result;

[0015] The result verification module is used to verify the obstacle point cloud detection results according to the obstacle space detection results.

[0016] In a third aspect, an embodiment of the present invention further provides a radar result detection device, the radar result detection device comprising:

[0017] at least one processor; and

[0018] a memory communicatively connected to at least one processor; wherein,

[0019] The memory stores a computer program that can be executed by at least one processor. The computer program is executed by at least one processor so that the at least one processor can execute the radar result detection method according to any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the radar result detection method of any embodiment of the present invention when executed.

[0021] The technical solution of the embodiment of the present invention obtains the obstacle point cloud detection result in the detection area of ​​the vehicle to be detected sent by the radar, performs target detection on the image of the acquired identifiable area, obtains the obstacle image detection result, determines the obstacle space detection result according to the obstacle image detection result, verifies the obstacle point cloud detection result according to the obstacle space detection result, and can automatically verify the obstacle point cloud detection result, thereby improving the accuracy of radar result detection.

[0022] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 is a flow chart of a radar result detection method provided according to Embodiment 1 of the present invention;

[0025] Figure 2 is a flow chart of a radar result detection method provided according to Embodiment 2 of the present invention;

[0026] Figure 3is a schematic diagram of a radar result detection area provided according to Embodiment 1 of the present invention;

[0027] Figure 4 is a schematic diagram of a radar result detection area provided according to Embodiment 2 of the present invention;

[0028] Figure 5 is a structural diagram of a radar result detection device provided according to an embodiment of the present invention;

[0029] Figure 6 It is a schematic diagram of the structure of a radar result detection device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] In the technical solution of the embodiment of the present invention, the acquisition, storage and application of the obstacle point cloud detection results involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0033] Embodiment 1

[0034] Figure 1 The flowchart of a radar result detection method provided in the first embodiment of the present invention is shown in FIG. The embodiment of the present invention is applicable to the case of radar result detection, and the method can be executed by a radar result detection device, and the radar result detection device can be implemented in the form of hardware and / or software.

[0035] See also Figure 1The radar result detection method shown includes:

[0036] S101, obtaining obstacle point cloud detection results in a to-be-detected area of ​​a to-be-detected vehicle sent by a radar.

[0037] The vehicle to be detected may be a vehicle to be subjected to obstacle detection. The area to be detected may be an area to be subjected to obstacle detection corresponding to the vehicle to be detected. The obstacle point cloud detection result may be a collection of data of obstacles detected by the vehicle to be detected in the area to be detected.

[0038] Specifically, a point cloud refers to a set of discrete three-dimensional spatial data points obtained by a radar system. The radar system uses electromagnetic waves or lasers to transmit signals, and measures the distance, position, shape and other information of the target object by receiving the echo of the signal. When the radar system scans the surrounding environment, it will obtain a series of echo signals, which are processed to form a set of discrete three-dimensional coordinate points, namely, a point cloud. Each point in the point cloud contains the coordinate information of a certain position in space, usually in the form of (x, y, z), indicating the three-dimensional position of the point. In addition, each point in the point cloud may also carry other attribute information, such as intensity value, reflectivity, color, etc. The radar can detect obstacles in all directions around the vehicle to be detected. The obstacle point cloud detection results of the vehicle to be detected in the area to be detected sent by the radar can be obtained by limiting the area to be detected, thereby reducing the amount of data to be analyzed.

[0039] In one example, the obstacle point cloud detection results in the detection area of ​​the vehicle to be detected sent by the radar are obtained by a depth-first traversal algorithm. The obstacle point cloud detection results include: car 1-(1,1,1)-(0,0,0)(2,2,2); car 2-(4,4,4)-(3,3,3)(5,5,5).

[0040] S102: Perform target detection on the acquired image of the identifiable area to obtain an obstacle image detection result.

[0041] Among them, target detection can be detecting target objects from a given image or video, determining their position and size, and classifying or identifying them at the same time. The identifiable area can be the area to be detected. The obstacle image detection result can be a collection of data of obstacles identified in the identifiable area.

[0042] Specifically, image data collected by a camera in the identifiable area may be obtained, and target detection may be performed on the image data to obtain an obstacle image detection result in the identifiable area. The target detection algorithm may be Harr+Adaboost, HOG+SVM, DPM, R-CNN, Fast R-CNN, or YOLO, etc., and the embodiment of the present invention is not limited thereto.

[0043] In an example, in a recognizable area, image data of a camera device is obtained, and target detection is performed on the image data using the YOLO algorithm to obtain obstacle image detection results, which include: car 1 - (1,1,1) - (0,0,0) (2,2,2); car 2 - (4,4,4) - (3,3,3) (5,5,5).

[0044] S103: Determine an obstacle space detection result according to the obstacle image detection result.

[0045] The obstacle space detection result may be coordinate data obtained by converting coordinate data in the obstacle image detection result of target detection into coordinate data.

[0046] Specifically, when performing target detection on image data, the coordinate data of the obstacle corresponds to the vehicle body coordinate system. However, the point cloud data is recorded in the laser radar coordinate system, so it is necessary to obtain the conversion matrix between the vehicle body coordinate system and the laser radar coordinate system, convert the coordinate data corresponding to the obstacle in the vehicle body coordinate system to the laser radar coordinate system, and determine the obstacle space detection result.

[0047] In an example, the coordinates of the center point of obstacle 1 detected in the vehicle body coordinate system are (x, y, z), and the transformation matrix between the vehicle body coordinate system and the laser radar coordinate system is matrix 1. Multiplying (x, y, z) with matrix 1, the coordinates of the center point of obstacle 1 in the laser radar coordinate system (x ′ ,y ′ ,z′).

[0048]

[0049] S104: Verify the obstacle point cloud detection result according to the obstacle space detection result.

[0050] Specifically, the obstacle point cloud detection result is evaluated and verified according to the obstacle space detection result obtained through target detection, the obstacle space detection result is used as the true value, the obstacle point cloud detection result is used as the test value, and the accuracy of the obstacle point cloud detection result is calculated.

[0051] In one example, the obstacle point cloud detection results include: car 1-(1,1,1)-(0,0,0)(2,2,2); car 2-(4,4,4)-(3,3,3)(5,5,5). The obstacle space detection results include: car 1-(1,1,1)-(0,0,0)(2,2,2); car 2-(4,4,4)-(3,3,3)(5,5,5). The obstacle point cloud detection results are verified according to the obstacle space detection results, and the accuracy of the obstacle point cloud detection results is 100%.

[0052] The technical solution of the embodiment of the present invention obtains the obstacle point cloud detection result in the detection area of ​​the vehicle to be detected sent by the radar, performs target detection on the image of the acquired identifiable area, obtains the obstacle image detection result, determines the obstacle space detection result according to the obstacle image detection result, verifies the obstacle point cloud detection result according to the obstacle space detection result, can automatically verify the obstacle point cloud detection result, improves the accuracy of radar result detection, and can automatically verify the obstacle point cloud detection result, improves the accuracy of radar detection result.

[0053] Embodiment 2

[0054] Figure 2 This is a flow chart of a radar result detection method provided in Embodiment 2 of the present invention. Based on the above embodiments, this embodiment of the present invention optimizes and improves the radar result detection operation.

[0055] Further, "verifying the obstacle point cloud detection results according to the obstacle space detection results" is refined into "the obstacle point cloud detection results include: the obstacle type, obstacle center coordinates and detection frame area of ​​at least one obstacle; the obstacle image detection results include: the object type, object center coordinates and true value frame area of ​​at least one target object; for each obstacle, the obstacle center coordinates in the obstacle point cloud detection results are compared with the object center coordinates corresponding to each target object to determine the object at the same position corresponding to the obstacle; the obstacle type corresponding to the obstacle is compared with the object type corresponding to the object at the same position to determine the type comparison result; the obstacle space detection result is calculated according to the type comparison result, the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position" to improve the operation of radar result detection.

[0056] It should be noted that for the parts not described in detail in the embodiments of the present invention, reference may be made to the descriptions of other embodiments.

[0057] See also Figure 2 The radar result detection method shown includes:

[0058] S201, obtaining obstacle point cloud detection results in the detection area of ​​the vehicle to be detected sent by the radar.

[0059] S202: Perform target detection on the acquired image of the identifiable area to obtain an obstacle image detection result.

[0060] S203: Determine an obstacle space detection result according to the obstacle image detection result.

[0061] S204. The obstacle point cloud detection result includes: the obstacle type, obstacle center coordinates and detection frame area of ​​at least one obstacle; the obstacle image detection result includes: the object type, object center coordinates and true value frame area of ​​at least one target object; for each obstacle, the obstacle center coordinates in the obstacle point cloud detection result are compared with the object center coordinates corresponding to each target object to determine the same-position object corresponding to the obstacle.

[0062] Among them, the obstacle type may be the type of obstacle detected by the radar. The obstacle center coordinates may be the coordinates of the center point of the obstacle detected by the radar. The detection frame area may be the area selected by the detection frame at the location of the obstacle detected by the radar from the driver's perspective. The target object may be an obstacle obtained by target detection. The object type may be the type of obstacle detected by the target. The object center coordinates may be the coordinates of the center point of the obstacle detected by the target. The true value frame area may be the area selected by the detection frame at the location of the obstacle detected by the target from the driver's perspective. The same-position object is used to describe that the coordinates of the center points of the obstacle detected by the radar and the obstacle detected by the target are the same.

[0063] Specifically, at least one obstacle is detected by radar, and for each obstacle, the coordinates of the center point corresponding to each obstacle are obtained. The coordinates of the center point corresponding to at least one target object obtained by target detection are obtained. For each obstacle, the obstacle center coordinates corresponding to the obstacle are compared with the object center coordinates corresponding to each target object, and the target object at the same position as each obstacle is obtained.

[0064] In one example, the obstacle point cloud detection results include: car 1-(1,1,1)-(0,0,0)(2,2,2); car 2-(4,4,4)-(3,3,3)(5,5,5). The obstacle space detection results include: car 3-(1,1,1)-(0,0,0)(2,2,2); truck 4-(8,8,8)-(7,7,7)(9,9,9). For car 1 and car 2, the obstacle center coordinates (1,1,1) of car 1 and the obstacle center coordinates (4,4,4) of car 2 in the obstacle point cloud detection results are compared with the object center coordinates corresponding to each target object, and it is determined that the object at the same location corresponding to car 1 is car 3.

[0065] S205: Compare the obstacle type corresponding to the obstacle with the object type corresponding to the object at the same position to determine a type comparison result.

[0066] The type comparison result may be a result of comparing the obstacle type of the obstacle with the object type corresponding to the object at the same position.

[0067] Specifically, the obstacle type and the object type with the same center point coordinates are compared to determine the type comparison result. If the type comparison result is the same type, it indicates that the position and type of the obstacle detected by the radar are correct; if the type comparison result is different types, it indicates that the position or type of the obstacle detected by the radar is incorrect.

[0068] In one example, the obstacle point cloud detection result includes: car 1-(1,1,1)-(0,0,0)(2,2,2). The obstacle space detection result includes: car 3-(1,1,1)-(0,0,0)(2,2,2). For car 1 and car 2, the object at the same location corresponding to car 1 is car 3. The obstacle type car corresponding to car 1 is compared with the object type car corresponding to car 3 at the same location, and it is determined that the type comparison result is the same type.

[0069] S206 , verifying the obstacle point cloud detection result according to the type comparison result, the detection frame area corresponding to the obstacle, and the true value frame area corresponding to the object at the same position.

[0070] Specifically, by comparing the center coordinates of the obstacle with the center coordinates of the target object, obstacles at the same position can be matched one-to-one to determine whether the spatial positions of the obstacles detected by the radar and the target are the same. The obstacle type of the obstacle at the same position is compared with the object type of the target object to determine whether the obstacle type detected by the radar is the same as the object type of the target object at the same position, and to determine whether the type of obstacle detected by the radar is accurate. The obstacle point cloud detection result is verified based on the calculation results of the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position, and the size of the intersection area of ​​the detection frame area of ​​the obstacle detected by the radar and the area of ​​the true value frame area corresponding to the target object is calculated to determine whether the size of the obstacle identified by the radar is correct.

[0071] In one example, the obstacle point cloud detection results include: car 1-(1,1,1)-(0,0,0)(2,2,2). The obstacle space detection results include: car 3-(1,1,1)-(0,0,0)(2,2,2). For car 1 and car 2, the object at the same position corresponding to car 1 is car 3. The obstacle type car corresponding to car 1 is compared with the object type car corresponding to car 3 at the same position, and the type comparison result is determined to be the same type. The ratio of the intersection of the detection frame area corresponding to car 1 and the truth frame area corresponding to car 3 at the same position to the detection frame area corresponding to car 1 is 80%, and it is determined that the obstacle point cloud detection result is accurate.

[0072] The embodiment of the present invention compares the center coordinates of the obstacle in the obstacle point cloud detection result with the object center coordinates corresponding to each target object for each obstacle, determines the object at the same position corresponding to the obstacle, compares the obstacle type corresponding to the obstacle with the object type corresponding to the object at the same position to determine the type comparison result, verifies the obstacle point cloud detection result according to the type comparison result, the detection frame area corresponding to the obstacle, and the true value frame area corresponding to the object at the same position, and can automatically verify the obstacle point cloud detection result through multi-dimensional indicators, thereby improving the accuracy of radar result detection.

[0073] Optionally, the obstacle point cloud detection result is verified according to the type comparison result, the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position, including: if the type comparison result is that the type is different, re-obtaining the obstacle point cloud detection result, and then verifying the obstacle point cloud detection result; if the type comparison result is that the type is the same, verifying the obstacle point cloud detection result according to the ratio of the intersection area of ​​the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position to the detection frame area corresponding to the obstacle.

[0074] Specifically, if the type comparison result is a different type, it means that the radar detection is inaccurate, and the obstacle point cloud detection result is re-acquired and then verified; if the type comparison result is the same type, it means that the radar detection is accurate, and the obstacle point cloud detection result is verified according to the ratio of the intersection area of ​​the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position to the detection frame area corresponding to the obstacle to determine whether the obstacle size recognition is accurate.

[0075] By performing different operations on different types of comparison results, the operation of radar result detection is refined and the accuracy of radar result detection is improved.

[0076] Optionally, after determining the obstacle space detection result according to the obstacle image detection result, the method further includes: counting the number of obstacles of the same type as the type comparison result to obtain the accurate identification number; the same type of obstacles refers to obstacles whose obstacle center coordinates are the same as the object center coordinates of the target object, and whose obstacle type is the same as the object type of the target object; counting the number of each obstacle to obtain the total number of obstacles; calculating the segmentation accuracy based on the accurate identification number and the total number of obstacles, and verifying the obstacle point cloud detection result based on the segmentation accuracy.

[0077] The accurate number of obstacles identified can be the accurate number of obstacles detected by the radar by comparing the obstacle point cloud detection result with the obstacle space detection result. The total number of obstacles can be the total number of obstacles detected by the radar in the area to be detected. The segmentation accuracy can be the accuracy of the radar in detecting obstacles.

[0078] Specifically, the number of obstacles with the same type as the type comparison result is counted to obtain the accurate number of identification, and then the number of obstacles detected by the radar is counted to obtain the total number of obstacles. The ratio of the accurate number of identification and the total number of obstacles is calculated to obtain the segmentation accuracy. The obstacle point cloud detection result can be verified according to the segmentation accuracy.

[0079] In one example, the number of obstacles with the same type compared is counted, and the number of accurately identified obstacles is 5. The number of obstacles detected by the radar is counted, and the total number of obstacles is 6. The ratio of the accurately identified number to the total number of obstacles is calculated to be 83%.

[0080] The segmentation accuracy is obtained by calculation, and the obstacle point cloud detection results are verified according to the segmentation accuracy, without the need for manual calculation, which improves the efficiency of radar result detection.

[0081] Optionally, after determining the obstacle space detection result according to the obstacle image detection result, the method further includes: identifying each target object to obtain at least one in-lane object and at least one non-in-lane object; counting the number of in-lane objects and non-in-lane objects to obtain the number of in-lane objects and the number of non-in-lane objects; calculating the segmentation recall rate according to the accurate number of identifications, the total number of obstacles, the number of in-lane objects and the number of non-in-lane objects; and verifying the obstacle point cloud detection result according to the segmentation recall rate.

[0082] Among them, the in-lane object may be an object located on the lane in the identifiable area. The non-in-lane object may be an object on the lane that is not in the identifiable area. The in-lane number may be the number of obstacles detected in the in-lane area. The non-in-lane number may be the number of obstacles detected in the identifiable area that are not in the in-lane area. The segmentation recall rate may be the ratio of the number of identified obstacles to the number of obstacles actually existing.

[0083] In one example, the radar detects 10 obstacles. The image of the identifiable area is subjected to target detection, and 7 obstacles are detected, including 5 objects in the lane and 2 objects outside the lane. The ratio of the number of objects in the lane to the total number of obstacles is 5 divided by 10, so the recall rate for the objects in the lane is 50%.

[0084] By identifying each target object, at least one in-lane object and at least one out-of-lane object are obtained, and the number of in-lane objects and out-of-lane objects is counted to obtain the number in the lane and the number out of the lane. The segmentation recall rate is calculated based on the accurate number of identifications, the total number of obstacles, the number in the lane and the number out of the lane. The obstacle point cloud detection result is verified according to the segmentation recall rate. The obstacle point cloud detection result can be verified by the segmentation recall rate, which improves the accuracy of radar result detection.

[0085] Optionally, each target object is identified to obtain at least one in-lane object and at least one out-of-lane object, including: obtaining a true value frame area corresponding to each target object; for each target object, comparing the true value frame area corresponding to the target object with the area to be detected to obtain an area overlap ratio corresponding to the target object; obtaining an area ratio threshold; for each target object, if the area overlap ratio is greater than or equal to the area ratio threshold, determining the target object as an in-lane object; if the area overlap ratio is less than the area ratio threshold, determining the target object as an out-of-lane object.

[0086] The area overlap ratio may be a ratio of the intersection area of ​​the true value frame area corresponding to the target object and the area to be detected to the area to be detected. The area ratio threshold may be a preset threshold of the area overlap ratio.

[0087] In an example, the true value frame area (0,0,0)(2,2,2) corresponding to the target object 1 is obtained; for the target object 1, the true value frame area (0,0,0)(2,2,2) corresponding to the target object 1 is compared with the area to be detected (0,0,0)(2,2,2), and the area overlap ratio corresponding to the target object 1 is 100%; the area ratio threshold of 80% is obtained; for the target object 1, if the area overlap ratio is greater than the area ratio threshold, the target object 1 is determined to be an object in the lane.

[0088] For each target object, if the area overlap ratio is greater than or equal to the area ratio threshold, the target object is determined to be an in-lane object; if the area overlap ratio is less than the area ratio threshold, the target object is determined to be a non-in-lane object. The target objects can be classified more finely, and different operations can be performed for different classifications to improve the accuracy of radar detection results.

[0089] Optionally, before obtaining the obstacle point cloud detection result in the detection area of ​​the vehicle to be detected sent by the radar, it also includes: obtaining map information, positioning information and target radius corresponding to the vehicle to be detected, the map information includes: lane area and intersection area; calculating the sensitive area corresponding to the vehicle to be detected according to the positioning information and target radius corresponding to the vehicle to be detected; and determining the intersection area of ​​the lane area, intersection area and sensitive area corresponding to the vehicle to be detected as the detection area corresponding to the vehicle to be detected.

[0090] The map information may be the road information where the vehicle to be detected is located. The positioning information may be the location data of the area to be detected. The target radius may be the radius of the circular area detectable by the vehicle to be detected. The lane area may be the area of ​​the lane in the map information of the area to be detected corresponding to the vehicle to be detected. The intersection area may be the area of ​​the intersection in the map information of the area to be detected corresponding to the vehicle to be detected. The sensible area may be the area formed by drawing a circle with the target radius with the location of the vehicle to be detected as the center.

[0091] Specifically, for the obstacle point cloud detection results, the obstacle point cloud detection results can be preprocessed, including: filtering the point cloud outside the road in combination with the positioning information and the map information, filtering the point cloud outside the detection range, and filtering the point cloud with abnormal values ​​(such as null values). Figure 3 As shown, the map provides lane areas and intersection areas. The lane areas and intersection areas within a certain range of the current position of the vehicle to be detected (such as the target radius R in the figure) are obtained based on the positioning information, and the intersection area of ​​the areas is used as the area to be detected. For the point cloud, ground detection is performed, and the ground point cloud is filtered out through the ground segmentation algorithm, and the obstacle point cloud above the ground is retained. The ground segmentation algorithm includes but is not limited to: ground plane fitting or ground line fitting under polar coordinates, etc., and the present invention does not make specific restrictions. According to the map information and positioning information, obstacles outside the map are filtered out, such as Figure 3 Obstacle A in the middle, keep the obstacles in the map such as Figure 4 For obstacles in the middle, if it is a motor vehicle, the number of vertices of the obstacle in the lane is determined. The obstacle has 8 vertices in total. If 6 or more are in the lane area, the target is retained. Figure 3 Obstacle C, otherwise it will not be retained. Figure 3 Obstacle D; if it is a non-motor vehicle, if it is at the road boundary, it will not be retained, such as Figure 3 Medium obstacle E.

[0092] Embodiment 3

[0093] Figure 5The structure diagram of a radar result detection device provided in Embodiment 3 of the present invention is shown in FIG. The embodiment of the present invention is applicable to the case of radar result detection, the device can execute the radar result detection method, and the device can be implemented in the form of hardware and / or software.

[0094] See also Figure 5 The radar result detection device shown includes: a point cloud data acquisition module 501, an image data acquisition module 502, a space data acquisition module 503 and a result verification module 504, wherein:

[0095] The point cloud data acquisition module 501 is used to acquire the obstacle point cloud detection result in the detection area of ​​the vehicle to be detected sent by the radar;

[0096] The image data acquisition module 502 is used to perform target detection on the acquired image of the identifiable area to obtain an obstacle image detection result;

[0097] The spatial data acquisition module 503 is used to determine the obstacle space detection result according to the obstacle image detection result;

[0098] The result verification module 504 is used to verify the obstacle point cloud detection result according to the obstacle space detection result.

[0099] The technical solution of the embodiment of the present invention obtains the obstacle point cloud detection result in the detection area of ​​the vehicle to be detected sent by the radar, performs target detection on the image of the acquired identifiable area, obtains the obstacle image detection result, determines the obstacle space detection result according to the obstacle image detection result, verifies the obstacle point cloud detection result according to the obstacle space detection result, and can automatically verify the obstacle point cloud detection result, thereby improving the accuracy of radar result detection.

[0100] Optionally, the result verification module 504 includes:

[0101] A coordinate comparison unit is used to compare the obstacle center coordinates in the obstacle point cloud detection result with the object center coordinates corresponding to each target object for each obstacle, and determine the object at the same position corresponding to the obstacle; the obstacle point cloud detection result includes: the obstacle type, obstacle center coordinates and detection frame area of ​​at least one obstacle; the obstacle image detection result includes: the object type, object center coordinates and true value frame area of ​​at least one target object;

[0102] A type comparison unit, used to compare the obstacle type corresponding to the obstacle with the object type corresponding to the object at the same position, and determine a type comparison result;

[0103] The area calculation unit is used to verify the obstacle point cloud detection result according to the type comparison result, the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position.

[0104] Optional, regional computing unit, specifically used for:

[0105] If the type comparison result is different, the obstacle point cloud detection result is obtained again, and then the obstacle point cloud detection result is verified;

[0106] If the type comparison result is the same type, the obstacle point cloud detection result is verified according to the ratio of the intersection area of ​​the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position to the detection frame area corresponding to the obstacle.

[0107] Optionally, the radar result detection device further includes:

[0108] The type statistics module is used to count the number of obstacles of the same type as the obstacle comparison result after determining the obstacle space detection result according to the obstacle image detection result, so as to obtain the accurate number of obstacles identified; the obstacles of the same type refer to obstacles whose obstacle center coordinates are the same as the object center coordinates of the target object, and whose obstacle type is the same as the object type of the target object;

[0109] The total number statistics module is used to count the number of each obstacle to obtain the total number of obstacles;

[0110] An accuracy calculation module is used to calculate the segmentation accuracy based on the accurate number of identified objects and the total number of obstacles;

[0111] The accuracy addition module is used to verify the obstacle point cloud detection results based on the segmentation accuracy.

[0112] Optionally, the radar result detection device further includes:

[0113] An object recognition module is used to recognize each target object and obtain at least one in-lane object and at least one out-lane object;

[0114] A lane data statistics module is used to count the number of objects in each lane and the number of objects outside the lane to obtain the number of objects in the lane and the number of objects outside the lane;

[0115] A recall calculation module is used to calculate the segmentation recall based on the number of accurate identifications, the total number of obstacles, the number of objects in the lane, and the number of objects not in the lane;

[0116] The recall rate module is added to verify the obstacle point cloud detection results based on the segmentation recall rate.

[0117] Optional object recognition module, specifically used for:

[0118] Get the true value box area corresponding to each target object;

[0119] For each target object, the true value frame area corresponding to the target object is compared with the area to be detected to obtain the area overlap ratio corresponding to the target object;

[0120] Get the area ratio threshold;

[0121] For each target object, if the area overlap ratio is greater than or equal to the area ratio threshold, the target object is determined to be an object within the lane;

[0122] If the area overlap ratio is less than the area ratio threshold, the target object is determined as an object outside the lane.

[0123] Optionally, the radar result detection device further includes:

[0124] The information acquisition module is used to obtain the map information, positioning information and target radius corresponding to the vehicle to be detected before obtaining the obstacle point cloud detection result in the detection area of ​​the vehicle to be detected sent by the radar, and the map information includes: lane area and intersection area;

[0125] A sensing area calculation module is used to calculate the sensing area corresponding to the vehicle to be detected based on the positioning information and target radius corresponding to the vehicle to be detected;

[0126] The intersection area determination module is used to determine the intersection area of ​​the lane area, the intersection area and the sensible area corresponding to the vehicle to be detected as the area to be detected corresponding to the vehicle to be detected.

[0127] The radar result detection device provided in the embodiment of the present invention can execute the radar result detection method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the radar result detection method.

[0128] Embodiment 4

[0129] Figure 6 A schematic structural diagram of a radar result detection device 600 that can be used to implement an embodiment of the present invention is shown.

[0130] like Figure 6As shown, the radar result detection device 600 includes at least one processor 601, and a memory connected to the at least one processor 601, such as a read-only memory (ROM) 602, a random access memory (RAM) 603, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 601 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 602 or the computer program loaded from the storage unit 608 to the random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the radar result detection device 600 can also be stored. The processor 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0131] Multiple components in the radar result detection device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the radar result detection device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0132] The processor 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 601 performs the various methods and processes described above, such as a radar result detection method.

[0133] In some embodiments, the radar result detection method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the radar result detection device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the processor 601, one or more steps of the radar result detection method described above may be performed. Alternatively, in other embodiments, the processor 601 may be configured to perform the radar result detection method in any other appropriate manner (e.g., by means of firmware).

[0134] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0135] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0136] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0137] To provide interaction with a user, the systems and techniques described herein may be implemented on a radar result detection device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which a user can provide input to the radar result detection device. Other types of devices may also be used to provide interaction with a user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0138] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0139] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS (Virtual Private Server) services.

[0140] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0141] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A radar result detection method, characterized in that: The method comprises: Obtaining the obstacle point cloud detection result in the detection area of ​​the vehicle to be detected sent by the radar; Performing target detection on the acquired image of the identifiable area to obtain obstacle image detection results; Determining an obstacle space detection result according to the obstacle image detection result; The obstacle point cloud detection result is verified according to the obstacle space detection result.

2. The method according to claim 1, characterized in that: The obstacle point cloud detection result includes: the obstacle type, obstacle center coordinates and detection frame area of ​​at least one obstacle; the obstacle image detection result includes: the object type, object center coordinates and true value frame area of ​​at least one target object; the checking of the obstacle point cloud detection result according to the obstacle space detection result includes: For each obstacle, compare the obstacle center coordinates in the obstacle point cloud detection result with the object center coordinates corresponding to each target object to determine the object at the same position corresponding to the obstacle; Comparing the obstacle type corresponding to the obstacle with the object type corresponding to the object at the same position to determine a type comparison result; The obstacle point cloud detection result is verified according to the type comparison result, the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position.

3. The method according to claim 2, characterized in that The verifying the obstacle point cloud detection result according to the type comparison result, the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position includes: If the type comparison result is that the types are different, re-obtaining the obstacle point cloud detection result, and then verifying the obstacle point cloud detection result; If the type comparison result is the same type, the obstacle point cloud detection result is verified according to the ratio of the intersection area of ​​the detection frame area corresponding to the obstacle and the true value frame area corresponding to the object at the same position to the detection frame area corresponding to the obstacle.

4. The method according to claim 2, characterized in that: After determining the obstacle space detection result according to the obstacle image detection result, the method further includes: The number of obstacles of the same type as the type comparison result is counted to obtain the accurate number of identification; the obstacles of the same type refer to obstacles whose obstacle center coordinates are the same as the object center coordinates of the target object, and whose obstacle type is the same as the object type of the target object; Counting the number of each obstacle to obtain the total number of obstacles; The segmentation accuracy is calculated based on the accurate number of identified objects and the total number of obstacles; The obstacle point cloud detection result is verified according to the segmentation accuracy.

5. The method according to claim 4, characterized in that After determining the obstacle space detection result according to the obstacle image detection result, the method further includes: Identify each of the target objects to obtain at least one in-lane object and at least one out-of-lane object; Counting the number of objects in each lane and the number of objects not in the lane to obtain the number of objects in the lane and the number of objects not in the lane; The segmentation recall rate is calculated according to the accurate number of recognitions, the total number of obstacles, the number in the lane, and the number out of the lane; The obstacle point cloud detection result is verified according to the segmentation recall rate.

6. The method according to claim 5, characterized in that The identifying of each of the target objects to obtain at least one in-lane object and at least one out-of-lane object includes: Obtaining the true value frame area corresponding to each of the target objects; For each of the target objects, the true value frame area corresponding to the target object is compared with the area to be detected to obtain the overlap ratio of the area corresponding to the target object; Get the area ratio threshold; For each of the target objects, if the area overlap ratio is greater than or equal to the area ratio threshold, the target object is determined as an object within the lane; If the area overlap ratio is less than the area ratio threshold, the target object is determined as a non-lane object.

7. The method according to claim 5, characterized in that Before obtaining the obstacle point cloud detection result in the detection area of ​​the vehicle to be detected sent by the radar, it also includes: Obtaining map information, positioning information and target radius corresponding to the vehicle to be detected, wherein the map information includes: lane area and intersection area; The sensing area corresponding to the vehicle to be detected is calculated according to the positioning information and target radius corresponding to the vehicle to be detected; An intersection area of ​​the lane area, the intersection area and the sensible area corresponding to the to-be-detected vehicle is determined as the to-be-detected area corresponding to the to-be-detected vehicle.

8. A radar result detection device, characterized in that: The device comprises: A point cloud data acquisition module is used to obtain the obstacle point cloud detection results in the detection area of ​​the vehicle to be detected sent by the radar; An image data acquisition module is used to perform target detection on the acquired image of the identifiable area to obtain an obstacle image detection result; A spatial data acquisition module, used to determine an obstacle space detection result according to the obstacle image detection result; The result verification module is used to verify the obstacle point cloud detection result according to the obstacle space detection result.

9. A radar result detection device, characterized in that: The radar result detection device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the radar result detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the radar result detection method according to any one of claims 1 to 7 when executed.

Images