Point cloud data detection result enhancement method and apparatus, computer device, and medium

Abstract

The application relates to a point cloud data detection result enhancement method and device, computer equipment and a medium. The method comprises the following steps: acquiring point cloud data of a detection area; inputting the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; wherein the detection model is obtained by mapping first sample camera data into a point cloud coordinate system, correlatively matching a detection result of first sample point cloud data and a detection result of the first sample camera data according to the similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to a correlatively matched result; the first sample camera data and the first sample point cloud data are data of the same detection area. The method can improve the accuracy of the obtained enhanced detection result of the point cloud data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of laser radar, in particular to a point cloud data detection result enhancement method and device, computer equipment and medium. BACKGROUND

[0002] Laser radar is a radar system that obtains information about a target, such as target distance, direction, height, speed, attitude, and even shape parameters, by transmitting a detection signal (laser beam) to the target, then comparing the received signal (target echo) reflected from the target with the transmitted signal, and making appropriate processing.

[0003] However, due to the installation of the laser radar, a blind area will occur when the laser radar detects the target, and the point cloud data cannot be completely collected. In addition, the received point cloud data is easily affected by special weather and spatial obstructions, etc., resulting in information distortion and information loss problems. Therefore, it is necessary to enhance the received point cloud data to improve the detection capability of the point cloud data. In the traditional technology, the point cloud data received by multiple laser radars is stacked together to enhance the point cloud data by increasing the density of the received point cloud data. However, since the point cloud data received by the laser radar is inaccurate under the influence of weather or obstruction, stacking the point cloud data received by multiple laser radars will also result in inaccurate stacked point cloud data.

[0004] Therefore, the traditional point cloud data detection result has the problem of low detection accuracy. SUMMARY

[0005] Therefore, it is necessary to provide a point cloud data detection result enhancement method, device, computer equipment and medium that can improve the accuracy of point cloud data detection results.

[0006] A point cloud data detection result enhancement method, the method comprising:

[0007] Obtaining point cloud data of a detection area;

[0008] Inputting the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; wherein the detection model is obtained by mapping first sample camera data into a point cloud coordinate system, associating and matching the detection result of the first sample point cloud data and the detection result of the first sample camera data according to the similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to the association and matching result; the first sample camera data and the first sample point cloud data are data of the same detection area.

[0009] In one of the embodiments, the training method of the detection model comprises:

[0010] According to the pre-established mapping relationship, the detection result of the first sample point cloud data is mapped to the first sample camera data to obtain a mapping result of the detection result of the first sample point cloud data;

[0011] The first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data is calculated;

[0012] According to the first similarity, the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data are associated and matched;

[0013] According to the associated matching result, the first sample point cloud data is labeled to obtain a label of the first sample point cloud data;

[0014] The initial detection model is trained according to the first sample point cloud data and the label of the first sample point cloud data to obtain the detection model.

[0015] In one of the embodiments, the training method of the detection model comprises:

[0016] According to the pre-established mapping relationship, the first sample camera data is mapped to a point cloud data coordinate system to obtain a mapping result of the first sample camera data;

[0017] The second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data is calculated;

[0018] According to the second similarity, the detection result of the first sample point cloud data and the mapping result of the first sample camera data are associated and matched;

[0019] According to the associated matching result, the first sample point cloud data is labeled to obtain a label of the first sample point cloud data;

[0020] The initial detection model is trained according to the first sample point cloud data and the label of the first sample point cloud data to obtain the detection model.

[0021] In one of the embodiments, the method further comprises:

[0022] Obtaining second sample point cloud data;

[0023] Convert the second sample point cloud data to a camera coordinate system through an external conversion matrix of the laser radar to the camera coordinate system to obtain processed second sample point cloud data; the processed second sample point cloud data is filtered out of points not in the camera collection angle and the laser radar is not received successfully.

[0024] Convert the processed second sample point cloud data to a polar coordinate system to obtain the polar angle of the processed second sample point cloud data.

[0025] Add a filling point to two points with equal polar angles of the processed second sample point cloud data by bisecting the polar length to obtain filled second sample point cloud data.

[0026] Convert the filled second sample point cloud data from a point cloud coordinate system to a pixel coordinate system to establish a mapping relationship pair of a depth point of the filled second sample point cloud data and a pixel point in the pixel coordinate system.

[0027] According to the mapping relationship pair of the depth point of the filled second sample point cloud data and the pixel point in the pixel coordinate system, fit the pixel coordinate information corresponding to the filled second sample point cloud data to obtain a calibration parameter.

[0028] Match third sample point cloud data and third sample camera data of the same sample detection area using the calibration parameter.

[0029] According to the matching results of the third sample point cloud data and the third sample camera data, establish the mapping relationship.

[0030] A point cloud data detection result enhancement method, the method comprising:

[0031] Obtain point cloud data and camera data of a detection area.

[0032] Detect the point cloud data to obtain a detection result of the point cloud data.

[0033] Detect the camera data to obtain a detection result of the camera data.

[0034] According to a pre-established mapping relationship, associate and match the detection result of the point cloud data and the detection result of the camera data.

[0035] According to the associated matching result, enhance the detection result of the point cloud data to obtain an enhanced detection result of the point cloud data.

[0036] In one embodiment, the associated matching of the detection result of the point cloud data and the detection result of the camera data according to the pre-established mapping relationship comprises:

[0037] According to the mapping relationship, the detection result of the point cloud data is mapped to the camera data to obtain a mapping result of the detection result of the point cloud data;

[0038] A first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data is calculated;

[0039] According to the first similarity, the detection result of the point cloud data and the detection result of the camera data are associated and matched.

[0040] In one of the embodiments, the associated matching of the detection result of the point cloud data and the detection result of the camera data according to the pre-established mapping relationship comprises:

[0041] According to the mapping relationship, the camera data is mapped into a point cloud data coordinate system to obtain a mapping result of the camera data;

[0042] A second similarity between the mapping result of the camera data and the detection result of the point cloud data is calculated;

[0043] According to the second similarity, the detection result of the point cloud data and the detection result of the camera data are associated and matched.

[0044] In one of the embodiments, the enhancement of the detection result of the point cloud data according to the associated matching result to obtain an enhanced detection result of the point cloud data comprises:

[0045] If the associated matching result is that the detection result of the point cloud data and the detection result of the camera data are not matched successfully, the detection result of the camera data is determined as the enhanced detection result of the point cloud data.

[0046] In one of the embodiments, the method further comprises:

[0047] According to the positioning frame of the detection result of the camera data, a positioning point of the detection result of the camera data in a point cloud coordinate system is determined;

[0048] According to the positioning point and a pre-established fitting positioning relationship, a position of the detection result of the camera data in the point cloud coordinate system is determined;

[0049] The detection result of the camera data is associated and matched by using a video tracking algorithm and the position of the detection result of the camera data in the point cloud coordinate system to obtain a driving track of the detection result of the camera data in the point cloud coordinate system;

[0050] According to the detection result of the camera data, a driving track in the point cloud coordinate system is obtained, and feature information of the detection result of the camera data is obtained; the feature information of the detection result of the camera data includes at least one of position, speed, heading angle, and acceleration of the detection result of the camera data.

[0051] In one of the embodiments, the method further comprises:

[0052] According to the size of the bounding box of the detection result of the camera data, the size of the detection result of the camera data is obtained.

[0053] In one of the embodiments, the method further comprises:

[0054] Obtaining second sample point cloud data;

[0055] Converting the second sample point cloud data to a camera coordinate system through an external conversion matrix of the laser radar to the camera coordinate system to obtain processed second sample point cloud data; the processed second sample point cloud data is filtered out of points not within the camera collection angle and points not successfully received by the laser radar;

[0056] Converting the processed second sample point cloud data to a polar coordinate system to obtain a polar angle of the processed second sample point cloud data;

[0057] Adding a filling point to two points with equal polar angles of the processed second sample point cloud data through equal division of the polar length to obtain filled second sample point cloud data;

[0058] Converting the filled second sample point cloud data from the point cloud coordinate system to a pixel coordinate system to establish a mapping relationship pair between a depth point of the filled second sample point cloud data and a pixel point in the pixel coordinate system;

[0059] According to the mapping relationship pair between the depth point of the filled second sample point cloud data and the pixel point in the pixel coordinate system, pixel coordinate information corresponding to the filled second sample point cloud data is fitted to obtain a calibration parameter;

[0060] Using the calibration parameter, third sample point cloud data and third sample camera data of the same sample detection region are matched;

[0061] According to the matching result of the third sample point cloud data and the third sample camera data, the mapping relationship is established.

[0062] An enhancement device for point cloud data detection results, the device comprising:

[0063] An acquisition module is configured to acquire point cloud data of a detection region.

[0064] An enhancement module is configured to input the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; wherein the detection model is obtained by mapping first sample camera data into a point cloud coordinate system, correlating and matching a detection result of first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to a correlation matching result.

[0065] An enhancement device for a point cloud data detection result, the device comprising:

[0066] A first acquisition module is configured to acquire point cloud data and camera data of a detection region.

[0067] A second acquisition module is configured to detect the point cloud data to obtain a detection result of the point cloud data.

[0068] A third acquisition module is configured to detect the camera data to obtain a detection result of the camera data.

[0069] A correlation module is configured to correlate and match the detection result of the point cloud data and the detection result of the camera data according to a preset mapping relationship.

[0070] An enhancement module is configured to enhance the detection result of the point cloud data according to the correlation matching result to obtain an enhanced detection result of the point cloud data.

[0071] A computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the following steps when executing the computer program:

[0072] Acquiring point cloud data of a detection region.

[0073] Inputting the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; wherein the detection model is obtained by mapping first sample camera data into a point cloud coordinate system, correlating and matching a detection result of first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to a correlation matching result.

[0074] A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the following steps:

[0075] Obtaining point cloud data of a detection area;

[0076] Inputting the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; wherein the detection model is obtained by mapping first sample camera data into a point cloud coordinate system, correlatively matching a detection result of the first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to a correlatively matching result.

[0077] The point cloud data detection result enhancement method, device, computer device and medium have the following advantages. The detection model is obtained by mapping first sample camera data in first sample camera data and first sample point cloud data of a same detection area into a point cloud coordinate system, correlatively matching a detection result of the first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to a correlatively matching result of the detection result of the first sample point cloud data and the detection result of the first sample camera data. The detection model obtained in this way can accurately detect input point cloud data, solves the problem of inaccurate detection results caused by point cloud data information distortion and information loss due to occlusion, weather and other reasons, and, since the information collected by the camera data is relatively complete, the accuracy of the detection result of the camera data is high, the correlatively matching of the detection result of the first sample point cloud data and the detection result of the second sample camera data can be accurately performed according to the similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, the preset initial detection model can be accurately trained according to the correlatively matching result, and the accuracy of the enhanced detection result of the point cloud data obtained by the detection model is further improved. BRIEF DESCRIPTION OF DRAWINGS

[0078] Figure 1 An internal structure schematic diagram of a computer device provided for an embodiment;

[0079] Figure 2 A flowchart of the point cloud data detection result enhancement method in an embodiment;

[0080] Figure 3 A flowchart of the point cloud data detection result enhancement method in another embodiment;

[0081] Figure 4 Flowchart of the method for enhancing the detection result of the point cloud data in another embodiment;

[0082] Figure 5 Flowchart of the method for enhancing the detection result of the point cloud data in another embodiment;

[0083] Figure 5a Diagram of the second sample point cloud data in an embodiment;

[0084] Figure 5b Diagram of filtering out the point cloud data not within the camera's collection angle in an embodiment;

[0085] Figure 5c Diagram of filtering out the second sample point cloud data not within the camera's collection angle and not successfully received by the laser radar in an embodiment;

[0086] Figure 5d Flowchart of obtaining the polar angle of the processed second sample point cloud data in an embodiment;

[0087] Figure 5e Flowchart of obtaining the filled second sample point cloud data in an embodiment;

[0088] Figure 6 Flowchart of the method for enhancing the detection result of the point cloud data in an embodiment;

[0089] Figure 7 Flowchart of the method for enhancing the detection result of the point cloud data in another embodiment;

[0090] Figure 8 Flowchart of the method for enhancing the detection result of the point cloud data in another embodiment;

[0091] Figure 9 Flowchart of the method for enhancing the detection result of the point cloud data in another embodiment;

[0092] Figure 10 Flowchart of the method for enhancing the detection result of the point cloud data in another embodiment;

[0093] Figure 11 Structural block diagram of the device for enhancing the detection result of the point cloud data in an embodiment;

[0094] Figure 12 Structural block diagram of the device for enhancing the detection result of the point cloud data in an embodiment. DETAILED DESCRIPTION

[0095] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not intended to limit the present application.

[0096] The point cloud data detection result enhancement method provided by the embodiments of the present application can be applied to a computer device as shown in Figure 1 The computer device includes a processor and a memory connected through a system bus. The memory stores a computer program. When the processor executes the computer program, the steps of the method embodiments described below can be executed. Optionally, the computer device can further include a network interface, a display screen and an input device. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The network interface of the computer device is used to communicate with external terminals through network connection. Optionally, the computer device can be a server, a personal computer, a personal digital assistant, or other terminal devices such as a tablet computer, a mobile phone, etc. It can also be a cloud server or a remote server. The specific form of the computer device is not limited in the embodiments of the present application.

[0097] In one embodiment, as shown in Figure 2 , a point cloud data detection result enhancement method is provided. The method is applied to a computer device as shown in Figure 1 , including the following steps:

[0098] S201, obtaining point cloud data of a detection area.

[0099] The point cloud data is the data of the target object information recorded in the form of points by laser radar scanning. Each point contains three-dimensional coordinates. Specifically, the computer device obtains the point cloud data corresponding to the detection area. Optionally, the computer device can obtain the point cloud data corresponding to the detection area from the base station corresponding to the detection area. Optionally, the detection area can be an intersection or a station area. Optionally, the point cloud data of the detection area can be collected by any one of 8-line laser radar, 16-line laser radar, 24-line laser radar, 32-line laser radar, 64-line laser radar and 128-line laser radar.

[0100] S202, input the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; wherein the detection model is obtained by mapping first sample camera data into a point cloud coordinate system, correlatively matching a detection result of the first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to a correlatively matching result; the first sample camera data and the first sample point cloud data are data of a same detection region.

[0101] Specifically, the computer device inputs the above-obtained point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data. Wherein the preset detection model is obtained by mapping first sample camera data in first sample camera data and first sample point cloud data of a same detection region into a point cloud coordinate system, correlatively matching a detection result of the first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to a correlatively matching result of the detection result of the first sample point cloud data and the detection result of the first sample camera data. Optionally, the first sample point cloud data can be collected by any one of 8-line laser radar, 16-line laser radar, 24-line laser radar, 32-line laser radar, 64-line laser radar and 128-line laser radar, and the first sample camera data can be collected by any one of gun-type camera, half-sphere-type camera and sphere-type camera.

[0102] In the above enhancement method of the point cloud data detection result, since the model is obtained by mapping the first sample camera data in the first sample point cloud data to the point cloud coordinate system by the computer device, associating and matching the detection result of the first sample point cloud data and the detection result of the first sample camera data according to the similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training the preset initial detection model according to the associating and matching result of the detection result of the first sample point cloud data and the detection result of the first sample camera data, the detection model obtained in this way can accurately detect the input point cloud data, solves the problem of inaccurate detection result caused by point cloud data information distortion and information loss due to occlusion, weather, etc. In addition, since the information collected by the camera data is relatively complete, the accuracy of the detection result of the camera data is high, so the associating and matching of the detection result of the first sample point cloud data and the detection result of the second sample camera data can be accurately performed according to the similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, so that the preset initial detection model can be accurately trained according to the associating and matching result, and the accuracy of the enhanced detection result of the point cloud data obtained by the detection model is further improved.

[0103] In the above scene of inputting the point cloud data into the preset detection model to obtain the enhanced detection result of the point cloud data, the detection model needs to be trained in advance. In one embodiment, as shown in Figure 3 The training method of the above detection model comprises:

[0104] S301, according to the pre-established mapping relationship, mapping the detection result of the first sample point cloud data to the first sample camera data to obtain the mapping result of the detection result of the first sample point cloud data.

[0105] Specifically, the computer device maps the detection result of the first sample point cloud data to the first sample camera data according to the pre-established mapping relationship to obtain the mapping result of the detection result of the first sample point cloud data. It should be noted that the detection result of the first sample point cloud data includes the point cloud data wrapped by the target frame and the feature information corresponding to the point cloud data wrapped by the target frame. Optionally, the computer device can establish a mapping relationship of the detection result of the first sample point cloud data to the first sample camera data by using the calibration parameters of the first sample point cloud data and the first sample camera data, and map the detection result of the first sample point cloud data to the first sample camera data according to the mapping relationship to obtain the mapping result of the detection result of the first sample point cloud data.

[0106] S302, calculate a first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data.

[0107] Specifically, the computer device calculates a first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data. Optionally, the computer device can calculate an overlap degree between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data, and obtain the first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data according to the overlap degree between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data.

[0108] S303, according to the first similarity, the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data are associated and matched.

[0109] Specifically, the computer device associates and matches the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data according to the first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data. Optionally, the computer device can compare the first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data with a preset threshold (such as 0.6), and when the first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data reaches the preset threshold, it is considered that the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data are matched successfully, and a matching result of successful matching is obtained.

[0110] S304, according to the associated matching result, the first sample point cloud data is labeled to obtain a label of the first sample point cloud data.

[0111] Specifically, the computer device labels the first sample point cloud data according to the associated matching result of the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data, and obtains a label of the first sample point cloud data. For example, if the detection result of the first sample camera data is a car, the computer device maps the detection result of the first sample point cloud data to the first sample camera data, and the computer device labels the first sample point cloud data corresponding to the position of the car in the point cloud coordinate system as a car according to the position of the car detected by the first sample camera data, and obtains the label corresponding to the first sample point cloud data as a car.

[0112] S305, according to the first sample point cloud data and the label of the first sample point cloud data, the initial detection model is trained to obtain a detection model.

[0113] Specifically, the computer device trains the initial detection model according to the first sample point cloud data and the label of the first sample point cloud data, and obtains the detection model. It can be understood that the computer device can input the first sample point cloud data into the initial detection model to obtain the detection result of the first sample point cloud data, obtain the value of the loss function of the initial detection model according to the obtained detection result of the first sample point cloud data and the label of the first sample point cloud data, and then train the initial detection model according to the value of the loss function of the initial detection model until the value of the loss function of the initial detection model reaches a minimum value or a stable value to obtain the detection model.

[0114] In the embodiment, the computer device can accurately map the detection result of the first sample point cloud data to the first sample camera data according to the pre-established mapping relationship, improve the accuracy of the mapping result of the obtained detection result of the first sample point cloud data, and thus accurately calculate the first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data. Furthermore, the computer device can accurately associate and match the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data according to the first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data, obtain an associated matching result with high accuracy, and thus accurately label the first sample point cloud data according to the obtained associated matching result, improve the accuracy of the label of the obtained first sample point cloud data, and accurately train the initial detection model according to the first sample point cloud data and the label of the first sample point cloud data, thereby improving the accuracy of the obtained detection model.

[0115] In the above scenario of inputting the point cloud data into the preset detection model to obtain the enhanced detection result of the point cloud data, the detection model needs to be trained in advance. In one embodiment, as shown in Figure 4 the training method of the detection model includes:

[0116] S401, according to the pre-established mapping relationship, the first sample camera data is mapped into the point cloud data coordinate system to obtain the mapping result of the first sample camera data.

[0117] Specifically, the computer device maps the first sample camera data into a point cloud data coordinate system according to a pre-established mapping relationship, to obtain a mapping result of the first sample camera data. Optionally, the computer device can establish a mapping relationship of the first sample camera data mapping into the point cloud data coordinate system by using calibration parameters of the first sample point cloud data and the first sample camera data, and map the first sample camera data into the point cloud data coordinate system according to the mapping relationship, to obtain the mapping result of the first sample camera data.

[0118] S402, calculate a second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data.

[0119] Specifically, the computer device calculates a second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data. Optionally, the computer device can calculate an overlap degree between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and obtain the second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data according to the overlap degree between the mapping result of the first sample camera data and the detection result of the first sample point cloud data.

[0120] S403, according to the second similarity, perform associated matching on the detection result of the first sample point cloud data and the detection result of the first sample camera data.

[0121] Specifically, the computer device performs associated matching on the detection result of the first sample point cloud data and the detection result of the first sample camera data according to the second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data obtained above. Optionally, the computer device can compare the second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data with a preset threshold (such as 0.8), and when the second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data reaches the preset threshold, it is considered that the mapping result of the first sample camera data and the detection result of the first sample point cloud data are successfully matched, to obtain a matching result of successful matching.

[0122] S404, according to the associated matching result, label the first sample point cloud data, to obtain a label of the first sample point cloud data.

[0123] Specifically, the computer device labels the first sample point cloud data according to the associated matching result of the detection result of the first sample point cloud data and the mapping result of the first sample camera data, and obtains a label of the first sample point cloud data. For example, if the detection result of the first sample point cloud data is a car, the computer device maps the first sample camera data into the point cloud data coordinate system, and labels the first sample point cloud data at the corresponding position as a car according to the position of the car in the point cloud coordinate system detected by the first sample point cloud data, so that the label corresponding to the first sample point cloud data is a car.

[0124] In S405, the initial detection model is trained according to the first sample point cloud data and the label of the first sample point cloud data, and a detection model is obtained.

[0125] Specifically, the computer device trains the initial detection model according to the first sample point cloud data and the label of the first sample point cloud data, and obtains the detection model. It can be understood that the computer device can input the first sample point cloud data into the initial detection model to obtain the detection result of the first sample point cloud data, and obtain the value of the loss function of the initial detection model according to the detection result of the first sample point cloud data and the label of the first sample point cloud data. Then, the initial detection model is trained according to the value of the loss function of the initial detection model until the value of the loss function of the initial detection model reaches a minimum value or a stable value, and the detection model is obtained.

[0126] In this embodiment, the computer device can accurately map the first sample camera data into the point cloud data coordinate system according to the pre-established mapping relationship, improve the accuracy of the obtained mapping result of the first sample camera data, and accurately calculate the second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data. Furthermore, the detection result of the first sample point cloud data and the mapping result of the first sample camera data can be accurately associated and matched according to the second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and the associated matching result with high accuracy can be obtained. Therefore, the first sample point cloud data can be accurately labeled according to the obtained associated matching result, the accuracy of the obtained label of the first sample point cloud data is improved, and the initial detection model can be accurately trained according to the first sample point cloud data and the label of the first sample point cloud data, thereby improving the accuracy of the obtained detection model.

[0127] In the above, the detection result of the first sample point cloud data is mapped onto the first sample camera data according to the pre-established mapping relationship, or the first sample camera data is mapped into the scene in the point cloud data coordinate system according to the pre-established mapping relationship, and the mapping relationship needs to be established in advance. In one embodiment, for example, Figure 5As shown, the above method further comprises:

[0128] S501, acquiring second sample point cloud data.

[0129] Specifically, the computer device acquires second sample point cloud data. Optionally, the computer device can acquire the second sample point cloud data from a base station corresponding to a detection area. Optionally, the detection area can be a crossroads, a station area or the like. Optionally, the second sample point cloud data can be collected by any one of an 8-line laser radar, a 16-line laser radar, a 24-line laser radar, a 32-line laser radar, a 64-line laser radar, and a 128-line laser radar. For example, as shown in Figure 5a , Figure 5a is a schematic view of the acquired second sample point cloud data.

[0130] S502, converting the second sample point cloud data to a camera coordinate system through an external conversion matrix of the laser radar to the camera coordinate system, to obtain processed second sample point cloud data; the processed second sample point cloud data is filtered out of points not within the camera collection angle and points not successfully received by the laser radar.

[0131] Specifically, the computer device converts the above-mentioned second sample point cloud data to a camera coordinate system through an external conversion matrix of the laser radar to the camera coordinate system, to obtain processed second sample point cloud data. The processed second sample point cloud data is filtered out of points not within the camera collection angle and points not successfully received by the laser radar. For example, as shown in Figure 5b , Figure 5c , Figure 5b is filtered out of point cloud data not within the camera collection angle, Figure 5c is the processed second sample point cloud data filtered out of points not within the camera collection angle and points not successfully received by the laser radar.

[0132] S503, converting the processed second sample point cloud data to a polar coordinate system to obtain a polar angle of the processed second sample point cloud data.

[0133] Specifically, the computer device converts the above-mentioned processed second sample point cloud data to a polar coordinate system to obtain a polar angle of the processed second sample point cloud data. For example, as shown in Figure 5d , if the overhead point of the processed second sample point cloud data (X, Y, Z) is (X, Y, 0), the computer device converts the overhead point (X, Y) to a polar coordinate system on the plane of Z=0 to obtain a polar angle of the processed second sample point cloud data.

[0134] S504, adding a filling point to two points with equal polar angles of the processed second sample point cloud data by bisecting the polar length, to obtain filled second sample point cloud data.

[0135] Specifically, the computer device adds the filling points by bisecting the polar length for two points with equal polar angles of the processed second sample point cloud data, to obtain the filled second sample point cloud data. Exemplarily, as shown in FIG. 4, the laser radar has different line beams, and the computer device adds the filling points by bisecting the polar length for two points with equal polar angles and adjacent line beams, to obtain the filled second sample point cloud data. Figure 5e

[0136] S505, converting the filled second sample point cloud data from the point cloud coordinate system to the pixel coordinate system, to establish a mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system.

[0137] Specifically, the computer device converts the filled second sample point cloud data from the point cloud coordinate system to the pixel coordinate system, to establish a mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system. Optionally, the computer device can convert the filled second sample point cloud data from the point cloud coordinate system to the pixel coordinate system through an external conversion matrix of the laser radar to the camera coordinate system and an internal conversion matrix of the laser radar to the camera coordinate system, so that the three-dimensional coordinates of each filled second sample point cloud data and the coordinates mapped to the pixel coordinate system are in one-to-one correspondence, and thus the computer device can establish the mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system according to the mapping relationship pair.

[0138] S506, fitting the pixel coordinate information corresponding to the filled second sample point cloud data according to the mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system, to obtain the calibration parameter.

[0139] Specifically, the computer device fits the pixel coordinate information corresponding to the filled second sample point cloud data according to the mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system, to obtain the calibration parameter. Optionally, the computer device can fit the pixel coordinate information corresponding to the filled second sample point cloud data by using a circle, to obtain the calibration parameter.

[0140] S507, matching the third sample point cloud data and the third sample camera data of the same sample detection region by using the calibration parameter.

[0141] ​Specifically, the computer device matches the third sample point cloud data and the third sample camera data of the same sample detection area by using the calibration parameters obtained above. Optionally, the computer device can obtain the coordinates (x_0, y_0) of a pixel of a point in the third sample camera data, then obtain two circles closest to the point from the inside and outside according to the fitting circles of the line bundles described in S506 above, then obtain the intersection of the straight line x = x_0 and the two circles, take the mapping points (x_0, y_1), (x_0, y_2) of the two point clouds closest to the two circles in the line bundles corresponding to the two fitting circles, then obtain the actual point cloud coordinates (X_1, Y_1, Z_1), (X_2, Y_2, Z_2) corresponding to the mapping points by using the fitting parameters of the circles corresponding to the line bundles in S506 above, and then obtain the three-dimensional coordinates corresponding to (x_0, y_0) by proportionally obtaining (X_1, Y_1, Z_1), (X_2, Y_2, Z_2) from (x_0, y_0) to (x_0, y_1) and (x_0, y_2), at this time, the matching of the third sample point cloud data and the third sample camera data of the same sample detection area is completed.

[0142] S508, establishing a mapping relationship according to the matching result of the third sample point cloud data and the third sample camera data.

[0143] Specifically, the computer device establishes a mapping relationship between the detection result of the point cloud data and the camera data, or a mapping relationship between the camera data and the point cloud data coordinate system, according to the matching result of the third sample point cloud data and the third sample camera data.

[0144] In this embodiment, the computer device can convert the obtained second sample point cloud data to the camera coordinate system through the external conversion matrix of the laser radar to the camera coordinate system, accurately obtain the processed second sample point cloud data filtered out from the camera collection visual angle and the laser radar unsuccessful reception, and then convert the processed second sample point cloud data to the polar coordinate system to obtain the polar angle of the processed second sample point cloud data. Furthermore, the two points with equal polar angle of the processed second sample point cloud data can be added with filling points by equal division of the polar length to obtain the filled second sample point cloud data. Further, the filled second sample point cloud data can be converted from the point cloud coordinate system to the pixel coordinate system to establish the mapping relationship pair of the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system, so as to fit the pixel coordinate information corresponding to the filled second sample point cloud data to obtain the calibration parameter. Furthermore, the third sample point cloud data and the third sample camera data of the same sample detection region can be matched by using the calibration parameter, so that the mapping relationship between the detection result of the point cloud data and the camera data, or the mapping relationship between the camera data and the point cloud data coordinate system can be accurately established according to the matching result of the third sample point cloud data and the third sample camera data.

[0145] In one embodiment, as shown in Figure 6 , a point cloud data detection result enhancement method is provided. The method is applied to the computer device in Figure 1 for example, and includes the following steps:

[0146] S601, obtaining point cloud data and camera data of a detection region.

[0147] Specifically, the computer device obtains point cloud data and camera data of a detection region. Optionally, the computer device can obtain point cloud data and camera data corresponding to the detection region from a base station corresponding to the detection region. It should be noted that the point cloud data and camera data corresponding to the detection region are data of the same time and the same spatial dimension. Optionally, the detection region can be a crossroads, or a station area.

[0148] S602, detecting the point cloud data to obtain a detection result of the point cloud data.

[0149] Specifically, the computer device detects the obtained point cloud data of the detection area to obtain a detection result of the point cloud data. Optionally, the computer device can detect the point cloud data by using a preset deep learning model and a preset tracking algorithm. Optionally, the obtained detection result of the point cloud data can include information such as a position, a size, a speed, and a heading angle of the target object in a point cloud coordinate system. It can be understood that the point cloud data corresponding to the obtained detection area is inaccurate due to influences of a detection range, a detection blind area, and a weather environment of the laser radar, and thus the accuracy of the detection result of the point cloud data is affected. Therefore, the detection result of the point cloud data needs to be enhanced.

[0150] S603, detecting the camera data to obtain a detection result of the camera data.

[0151] Specifically, the computer device detects the obtained camera data to obtain a detection result of the camera data. Optionally, the computer device can detect the camera data by using a preset deep learning model and a preset tracking algorithm. Optionally, the obtained detection result of the camera data can include information such as a pixel rectangular frame of the target object, a target confidence, a target category, and a target tracking ID.

[0152] S604, according to a mapping relationship established in advance, associating and matching the detection result of the point cloud data and the detection result of the camera data.

[0153] Specifically, the computer device associates and matches the detection result of the point cloud data and the detection result of the camera data of the detection area according to the mapping relationship established in advance. Optionally, the computer device can establish a mapping relationship in which sample camera data is mapped into a point cloud data coordinate system by using calibration parameters of sample point cloud data and sample camera data, and associate and match the detection result of the point cloud data and the detection result of the camera data of the detection area according to the mapping relationship.

[0154] S605, according to the association and matching result, enhancing the detection result of the point cloud data to obtain an enhanced detection result of the point cloud data.

[0155] Specifically, the computer device enhances the detection result of the point cloud data of the detection area according to the association and matching result of the detection result of the point cloud data and the detection result of the camera data of the detection area to obtain an enhanced detection result of the point cloud data of the detection area. Optionally, the computer device can determine the detection result of the camera data as the enhanced detection result of the point cloud data according to the association and matching result of the detection result of the point cloud data and the detection result of the camera data of the detection area.

[0156] In the above enhancement method of the point cloud data detection result, the computer device can obtain the detection result of the point cloud data by detecting the point cloud data, and obtain the detection result of the camera data by detecting the camera data, so as to associate and match the detection result of the point cloud data and the detection result of the camera data according to the pre-established mapping relationship, and then enhance the detection result of the point cloud data of the detection region according to the association and matching result of the detection result of the point cloud data and the detection result of the camera data, thereby solving the problem of inaccurate detection result caused by information distortion and information loss of the point cloud data due to occlusion, weather and other reasons. In addition, since the information collected by the camera data is relatively complete, the accuracy of the detection result of the camera data is high, and the enhanced detection result of the point cloud data is obtained according to the association and matching result of the detection result of the point cloud data and the detection result of the camera data, so that the accuracy of the enhanced detection result of the point cloud data is improved.

[0157] In the above scene of associating and matching the detection result of the point cloud data and the detection result of the camera data according to the pre-established mapping relationship, in one embodiment, as shown in Figure 7 S604 includes:

[0158] S701, mapping the detection result of the point cloud data to the camera data according to the mapping relationship to obtain a mapping result of the detection result of the point cloud data.

[0159] Specifically, the computer device maps the detection result of the point cloud data to the camera data according to the above mapping relationship to obtain a mapping result of the detection result of the point cloud data. It should be noted that the detection result of the point cloud data includes point cloud data wrapped by a target box and feature information corresponding to the point cloud data wrapped by the target box. Optionally, the computer device can use the calibration parameters of the point cloud data and the camera data to establish a mapping relationship of the detection result of the point cloud data to the camera data, and map the detection result of the point cloud data to the camera data according to the mapping relationship to obtain the mapping result of the detection result of the point cloud data.

[0160] S702, calculating a first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data.

[0161] Specifically, the computer device calculates a first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data. Optionally, the computer device can calculate the overlap degree between the mapping result of the detection result of the point cloud data and the detection result of the camera data, and obtain the first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data according to the overlap degree between the mapping result of the detection result of the point cloud data and the detection result of the camera data.

[0162] S703, according to the first similarity, the detection result of the point cloud data and the detection result of the camera data are associated and matched.

[0163] Specifically, the computer device associates and matches the detection result of the point cloud data and the detection result of the camera data according to the first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data in the detection area. Optionally, the computer device can compare the first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data in the detection area with a preset threshold (such as 0.6). When the first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data reaches the preset threshold, it is considered that the detection result of the point cloud data and the detection result of the camera data are successfully matched, and a matching result of successful matching is obtained.

[0164] In this embodiment, the computer device can accurately map the detection result of the point cloud data of the detection area to the camera data according to the pre-established mapping relationship, improve the accuracy of the mapping result of the obtained detection result of the point cloud data, and thus accurately calculate the first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data, and further accurately associate and match the detection result of the point cloud data and the detection result of the camera data according to the first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data, and obtain a high-accuracy associated matching result.

[0165] In the above scenario of associating and matching the detection result of the point cloud data and the detection result of the camera data according to the pre-established mapping relationship, in one embodiment, as shown in Figure 8 S604 includes:

[0166] S801, according to the pre-established mapping relationship, the camera data is mapped into the point cloud data coordinate system to obtain the mapping result of the camera data.

[0167] Specifically, the computer device maps the camera data of the above detection area into the point cloud data coordinate system according to the pre-established mapping relationship to obtain the mapping result of the camera data. Optionally, the computer device can use the calibration parameters of the sample point cloud data and the sample camera data to establish a mapping relationship of the sample camera data mapped into the point cloud data coordinate system, and map the camera data into the point cloud data coordinate system according to the mapping relationship to obtain the mapping result of the camera data.

[0168] S802, calculate the second similarity between the mapping result of the camera data and the detection result of the point cloud data.

[0169] Specifically, the computer device calculates a second similarity between the mapping result of the camera data of the detection region and the detection result of the point cloud data. Optionally, the computer device can calculate an overlap degree between the mapping result of the camera data and the detection result of the point cloud data, and obtain the second similarity between the mapping result of the camera data and the detection result of the point cloud data according to the overlap degree between the mapping result of the camera data and the detection result of the point cloud data.

[0170] S803, according to the second similarity, the detection result of the point cloud data and the detection result of the camera data are associated and matched.

[0171] Specifically, the computer device associates and matches the detection result of the point cloud data and the detection result of the camera data according to the second similarity obtained between the mapping result of the camera data and the detection result of the point cloud data. Optionally, the computer device can compare the second similarity between the mapping result of the camera data and the detection result of the point cloud data with a preset threshold (such as 0.8), and when the second similarity between the mapping result of the camera data and the detection result of the point cloud data reaches the preset threshold, it is considered that the mapping result of the camera data and the detection result of the point cloud data are successfully matched, and a matching result of successful matching is obtained.

[0172] In this embodiment, the computer device can accurately map the camera data of the detection region into the point cloud data coordinate system according to the pre-established mapping relationship, improve the accuracy of the obtained mapping result of the camera data, and thus accurately calculate the second similarity between the mapping result of the camera data and the detection result of the point cloud data, and further accurately associate and match the detection result of the point cloud data and the detection result of the camera data according to the second similarity between the mapping result of the camera data and the detection result of the point cloud data, and obtain a high-accuracy associated matching result.

[0173] In the above scenario of enhancing the detection result of the point cloud data according to the associated matching result of the detection result of the point cloud data and the detection result of the camera data, obtaining the enhanced detection result of the point cloud data, in one embodiment, on the basis of the above embodiment, S605 includes: if the associated matching result is that the detection result of the point cloud data and the detection result of the camera data are not successfully matched, the detection result of the camera data is determined as the enhanced detection result of the point cloud data.

[0174] Specifically, if the result of the association matching between the detection result of the point cloud data and the detection result of the camera data is that the detection result of the point cloud data and the detection result of the camera data are not matched successfully, the computer device determines the detection result of the camera data in the detection region as the enhanced detection result of the point cloud data in the detection region. For example, if the detection result of the camera data in the detection region is a car, and the detection result of the point cloud data does not detect a target, after the computer device performs the association matching between the detection result of the camera data and the detection result of the point cloud data, the result is that the detection result of the camera data and the detection result of the point cloud data are not matched successfully, and then the computer device determines the detection result of the camera data as the enhanced detection result of the point cloud data.

[0175] In the embodiment, if the result of the association matching between the detection result of the point cloud data and the detection result of the camera data is that the detection result of the point cloud data and the detection result of the camera data are not matched successfully, the computer device determines the detection result of the camera data as the enhanced detection result of the point cloud data. Since the information collected by the camera data is relatively complete, the accuracy of the detection result of the camera data is relatively high, and thus, the detection result of the camera data is determined as the enhanced detection result of the point cloud data, which improves the accuracy of the enhanced detection result of the point cloud data.

[0176] In some scenarios, after the computer device obtains the detection result of the camera data, the computer device can further determine the feature information of the detection result of the camera data according to the detection result of the camera data. In one embodiment, as shown in FIG. 9, the method further includes: Figure 9

[0177] S901, determining a positioning point of the detection result of the camera data in the point cloud coordinate system according to the positioning frame of the detection result of the camera data.

[0178] Specifically, the computer device determines the positioning point of the detection result of the camera data in the point cloud coordinate system according to the positioning frame of the detection result of the camera data. Optionally, the computer device can determine the center point of the lower edge of the positioning frame of the detection result of the camera data as the positioning point of the detection result of the camera data in the point cloud coordinate system. Optionally, the computer device can also determine other points on the positioning frame of the detection result of the camera data as the positioning point of the detection result of the camera data in the point cloud coordinate system, for example, the center point of the positioning frame.

[0179] S902, determining the position of the detection result of the camera data in the point cloud coordinate system according to the positioning point and the pre-established fitting positioning relationship.

[0180] ​Specifically, the computer device determines the position of the detection result of the camera data in the point cloud coordinate system according to the determined positioning point of the detection result of the camera data in the point cloud coordinate system and the pre-established fitting positioning relationship. The pre-established fitting positioning relationship is a fitting relationship between the positioning point of the detection result of the camera data in the point cloud coordinate system and the position of the detection result of the camera data in the point cloud coordinate system. After the positioning point of the detection result of the camera data in the point cloud coordinate system is determined, the computer device can determine the position of the detection result of the camera data in the point cloud coordinate system according to the fitting positioning relationship.

[0181] S903, the detection result of the camera data is matched and associated by using the video tracking algorithm and the position of the detection result of the camera data in the point cloud coordinate system, to obtain a driving track of the detection result of the camera data in the point cloud coordinate system.

[0182] Specifically, the computer device matches and associates the detection result of the camera data by using the video tracking algorithm and the position of the detection result of the camera data in the point cloud coordinate system, to obtain a driving track of the detection result of the camera data in the point cloud coordinate system. Optionally, the computer device can predict the position of the detection result of the camera data in the point cloud coordinate system by using the video tracking algorithm, to obtain a driving track of the detection result of the camera data in the point cloud coordinate system, and then match and associate the position of the detection result of the camera data in the point cloud coordinate system with the detection result of the camera data, to obtain a driving track of the detection result of the camera data in the point cloud coordinate system.

[0183] S904, according to the driving track of the detection result of the camera data in the point cloud coordinate system, to obtain feature information of the detection result of the camera data; the feature information of the detection result of the camera data includes at least one of the position of the detection result of the camera data, the speed of the detection result of the camera data, the heading angle of the detection result of the camera data, and the acceleration of the detection result of the camera data.

[0184] Specifically, the computer device obtains the feature information of the detection result of the camera data according to the driving track of the detection result of the camera data in the point cloud coordinate system. The feature information of the detection result of the camera data includes at least one of the position of the detection result of the camera data, the speed of the detection result of the camera data, the heading angle of the detection result of the camera data, and the acceleration of the detection result of the camera data. It can be understood that the computer device can calculate the position of the detection result of the camera data, the speed of the detection result of the camera data, the heading angle of the detection result of the camera data, and the acceleration of the detection result of the camera data according to the driving track of the detection result of the camera data in the point cloud coordinate system, so as to obtain the feature information of the detection result of the camera data.

[0185] In this embodiment, the computer device can quickly determine the positioning point of the detection result of the camera data in the point cloud coordinate system according to the positioning box of the detection result of the camera data, so as to quickly determine the position of the detection result of the camera data in the point cloud coordinate system according to the positioning point of the detection result of the camera data in the point cloud coordinate system and the pre-established fitting positioning relationship, and further, the computer device can quickly obtain the driving track of the detection result of the camera data in the point cloud coordinate system by using the video tracking algorithm and the position of the detection result of the camera data in the point cloud coordinate system, and further, the computer device can quickly obtain the feature information of the detection result of the camera data according to the driving track of the detection result of the camera data in the point cloud coordinate system, thereby improving the efficiency of obtaining the feature information of the detection result of the camera data.

[0186] In some scenarios, the computer device also needs to determine the size of the detection result of the camera data according to the detection result of the camera data. In one embodiment, the above method further includes: obtaining the size of the detection result of the camera data according to the size of the positioning box of the detection result of the camera data.

[0187] Specifically, the computer device obtains the size of the detection result of the camera data according to the size of the positioning box of the detection result of the camera data. Optionally, the computer device can determine the size of the positioning box of the detection result of the camera data as the size of the detection result of the camera data. For example, if the size of the positioning box of the detection result of the camera data is 16*16, the computer device can determine the size of the detection result of the camera data as 16*16.

[0188] In this embodiment, the computer device can quickly obtain the size of the detection result of the camera data according to the size of the positioning box of the detection result of the camera data, thereby improving the efficiency of obtaining the size of the detection result of the camera data, and further, the computer device can accurately obtain the size of the detection result of the camera data according to the size of the positioning box of the detection result of the camera data, thereby improving the accuracy of the obtained size of the detection result of the camera data.

[0189] In the above scenario of mapping the detection result of the point cloud data to the camera data according to the pre-established mapping relationship, or mapping the camera data to the point cloud data coordinate system according to the pre-established mapping relationship, the mapping relationship needs to be pre-established. In one embodiment, as shown in Figure 10 the above method further includes:

[0190] S1001, obtaining second sample point cloud data.

[0191] Specifically, the computer device obtains second sample point cloud data. Optionally, the computer device can obtain the second sample point cloud data from a base station corresponding to a detection area. Optionally, the detection area can be an intersection or a station area. Optionally, the second sample point cloud data can be collected by any one of an 8-line laser radar, a 16-line laser radar, a 24-line laser radar, a 32-line laser radar, a 64-line laser radar, and a 128-line laser radar. For example, please continue to refer to Figure 5a , Figure 5a for a schematic diagram of the obtained second sample point cloud data.

[0192] S1002, converting the second sample point cloud data to a camera coordinate system through an external conversion matrix of the laser radar to the camera coordinate system, to obtain processed second sample point cloud data; the processed second sample point cloud data is filtered to remove points that are not within the camera collection angle and are not successfully received by the laser radar.

[0193] Specifically, the computer device converts the above-mentioned second sample point cloud data to a camera coordinate system through an external conversion matrix of the laser radar to the camera coordinate system, to obtain processed second sample point cloud data. The processed second sample point cloud data is filtered to remove points that are not within the camera collection angle and are not successfully received by the laser radar. For example, please continue to refer to Figure 5b 、 Figure 5c , Figure 5b for the point cloud data filtered to remove points that are not within the camera collection angle, Figure 5c for the processed second sample point cloud data filtered to remove points that are not within the camera collection angle and are not successfully received by the laser radar.

[0194] S1003, converting the processed second sample point cloud data to a polar coordinate system to obtain a polar angle of the processed second sample point cloud data.

[0195] Specifically, the computer device converts the above-mentioned processed second sample point cloud data to a polar coordinate system to obtain a polar angle of the processed second sample point cloud data. For example, please continue to refer to Figure 5d If the overhead point of the processed second sample point cloud data (X, Y, Z) is (X, Y, 0), the computer device converts the overhead point (X, Y) to a polar coordinate system on the plane of Z=0 to obtain a polar angle of the processed second sample point cloud data.

[0196] S1004, adding a filling point by bisecting the polar length for two points with equal polar angles of the processed second sample point cloud data, to obtain filled second sample point cloud data.

[0197] ​Specifically, the computer device adds a filling point to two points with equal polar angles of the processed second sample point cloud data by bisecting the polar length, to obtain the filled second sample point cloud data. For example, please continue to refer to Figure 5e The laser radar has different line beams, and the computer device adds a filling point to two points with equal polar angles and adjacent line beams by bisecting the polar length, to obtain the filled second sample point cloud data.

[0198] S1005, the computer device converts the filled second sample point cloud data from the point cloud coordinate system to the pixel coordinate system, to establish a mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system.

[0199] Specifically, the computer device converts the filled second sample point cloud data from the point cloud coordinate system to the pixel coordinate system, to establish a mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system. Optionally, the computer device can convert the filled second sample point cloud data from the point cloud coordinate system to the pixel coordinate system through an external conversion matrix of the laser radar to the camera coordinate system and an internal conversion matrix of the laser radar to the camera coordinate system. At this time, the three-dimensional coordinates of each filled second sample point cloud data and the coordinates mapped to the pixel coordinate system are in one-to-one correspondence, so that the computer device can establish a mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system according to the mapping relationship pair.

[0200] S1006, according to the mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system, the computer device fits the pixel coordinate information corresponding to the filled second sample point cloud data, to obtain the calibration parameter.

[0201] Specifically, the computer device fits the pixel coordinate information corresponding to the filled second sample point cloud data according to the mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system, to obtain the calibration parameter. Optionally, the computer device can fit the pixel coordinate information corresponding to the filled second sample point cloud data by using a circle, to obtain the calibration parameter.

[0202] S1007, the computer device matches the third sample point cloud data and the third sample camera data of the same sample detection region by using the calibration parameter.

[0203] Specifically, the computer device matches the third sample point cloud data and the third sample camera data of the same sample detection area by using the calibration parameters obtained above. Optionally, the computer device can obtain the coordinates (x_0, y_0) of a pixel of a point in the third sample camera data, then obtain two circles of the inner side and the outer side closest to the point according to the fitting circles of the line bundles described in S1006 above, then obtain the intersection points of the straight line x = x_0 and the two circles, take the mapping points (x_0, y_1), (x_0, y_2) of the two point clouds closest to the two circles in the line bundles corresponding to the two fitting circles, then obtain the actual point cloud coordinates (X_1, Y_1, Z_1), (X_2, Y_2, Z_2) corresponding to the mapping points by using the fitting parameters of the circles corresponding to the line bundles in S1006 above, and then obtain the three-dimensional coordinates corresponding to (x_0, y_0) by proportionally obtaining (X_1, Y_1, Z_1), (X_2, Y_2, Z_2) from (x_0, y_0) to (x_0, y_1) and (x_0, y_2). At this time, the matching of the third sample point cloud data and the third sample camera data of the same sample detection area is completed.

[0204] S1008, according to the matching result of the third sample point cloud data and the third sample camera data, a mapping relationship is established.

[0205] Specifically, the computer device establishes the mapping relationship between the detection result of the point cloud data and the camera data, or the mapping relationship between the camera data and the point cloud data coordinate system, according to the matching result of the third sample point cloud data and the third sample camera data.

[0206] In this embodiment, the computer device can convert the obtained second sample point cloud data to the camera coordinate system through the external conversion matrix of the laser radar to the camera coordinate system, accurately obtain the processed second sample point cloud data filtered out from the camera collection angle and the laser radar unsuccessful reception, and then convert the processed second sample point cloud data to the polar coordinate system to obtain the polar angle of the processed second sample point cloud data. Furthermore, the two points with equal polar angle of the processed second sample point cloud data can be added with filling points by equal division of the polar length to obtain the filled second sample point cloud data. Further, the filled second sample point cloud data can be converted from the point cloud coordinate system to the pixel coordinate system to establish the mapping relationship pair of the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system, so as to fit the pixel coordinate information corresponding to the filled second sample point cloud data to obtain the calibration parameter. Then, the third sample point cloud data and the third sample camera data of the same sample detection region can be matched by using the calibration parameter, so that the mapping relationship between the detection result of the point cloud data and the camera data, or the mapping relationship between the camera data and the point cloud data coordinate system can be accurately established according to the matching result of the third sample point cloud data and the third sample camera data.

[0207] It should be understood that, although Figure 2-10 the steps in the flowcharts are shown in sequence according to the arrows, these steps are not necessarily executed in sequence according to the arrows. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, Figure 2-10 at least part of the steps in the flowcharts can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times. The execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with other steps or steps or stages in other steps.

[0208] In one embodiment, as Figure 11 shown, a point cloud data detection result enhancement device is provided, comprising: a first acquisition module and an enhancement module, wherein:

[0209] The first acquisition module is configured to acquire point cloud data of a detection region.

[0210] The enhancement module is configured to input the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; the detection model is obtained by mapping first sample camera data into a point cloud coordinate system, associating and matching a detection result of the first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to an association matching result; the first sample camera data and the first sample point cloud data are data of a same detection region.

[0211] The point cloud data detection result enhancement device provided in the embodiment can execute the method embodiments, and has similar implementation principles and technical effects, which will not be described here.

[0212] On the basis of the above-mentioned embodiments, the device can further include a first mapping module, a first calculation module, a first matching module, a first labeling module and a first training module.

[0213] The first mapping module is configured to map the detection result of the first sample point cloud data to the first sample camera data according to a preset mapping relationship to obtain a mapping result of the detection result of the first sample point cloud data.

[0214] The first calculation module is configured to calculate a first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data.

[0215] The first matching module is configured to associate and match the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data according to the first similarity.

[0216] The first labeling module is configured to label the first sample point cloud data according to the association matching result to obtain a label of the first sample point cloud data.

[0217] The first training module is configured to train an initial detection model according to the first sample point cloud data and the label of the first sample point cloud data to obtain the detection model.

[0218] The point cloud data detection result enhancement device provided in the embodiment can execute the method embodiments, and has similar implementation principles and technical effects, which will not be described here.

[0219] On the basis of the above-mentioned embodiments, the device can further include a second mapping module, a second calculation module, a second matching module, a second labeling module and a second training module.

[0220] The second mapping module is configured to map the first sample camera data into a point cloud data coordinate system according to a pre-established mapping relationship to obtain a mapping result of the first sample camera data.

[0221] The second calculation module is configured to calculate a second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data.

[0222] The second matching module is configured to perform associated matching on the detection result of the first sample point cloud data and the mapping result of the first sample camera data according to the second similarity.

[0223] The second labeling module is configured to label the first sample point cloud data according to the associated matching result to obtain a label of the first sample point cloud data.

[0224] The second training module is configured to train the initial detection model according to the first sample point cloud data and the label of the first sample point cloud data to obtain the detection model.

[0225] The point cloud data detection result enhancement device provided in the embodiment can execute the method embodiments, and has similar implementation principles and technical effects, which will not be described here.

[0226] On the basis of the above-mentioned embodiments, the device can further include a second acquisition module, a third acquisition module, a fourth acquisition module, a fifth acquisition module, a first establishment module, a fitting module, a third matching module, and a second establishment module.

[0227] The second acquisition module is configured to acquire second sample point cloud data.

[0228] The third acquisition module is configured to convert the second sample point cloud data to a camera coordinate system through an external conversion matrix of the laser radar to the camera coordinate system to obtain processed second sample point cloud data; the processed second sample point cloud data is second sample point cloud data from which points that are not within the camera collection angle and points that are not successfully received by the laser radar are filtered out.

[0229] The fourth acquisition module is configured to convert the processed second sample point cloud data to a polar coordinate system to obtain a polar angle of the processed second sample point cloud data.

[0230] The fifth acquisition module is configured to add a filling point to two points with equal polar angles of the processed second sample point cloud data through equal division of a polar length to obtain filled second sample point cloud data.

[0231] The first establishment module is configured to convert the filled second sample point cloud data from a point cloud coordinate system to a pixel coordinate system to establish a mapping relationship pair of a depth point of the filled second sample point cloud data and a pixel point in the pixel coordinate system.

[0232] The fitting module is configured to fit pixel coordinate information corresponding to the filled second sample point cloud data according to the mapping relationship between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system, to obtain the calibration parameter.

[0233] The third matching module is configured to match the third sample point cloud data and the third sample camera data of the same sample detection region by using the calibration parameter.

[0234] The second establishing module is configured to establish the mapping relationship according to the matching result of the third sample point cloud data and the third sample camera data.

[0235] The point cloud data detection result enhancement device provided in this embodiment can execute the above-mentioned method embodiments, and has similar implementation principles and technical effects, which will not be described here again.

[0236] The specific limitations of the point cloud data detection result enhancement device can be referred to the limitations of the point cloud data detection result enhancement method described above, which will not be described here again. The modules in the point cloud data detection result enhancement device described above can be all or partially realized by software, hardware and combinations thereof. The modules described above can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the modules.

[0237] In one embodiment, as shown in Figure 12 A point cloud data detection result enhancement device is provided, which comprises a first obtaining module, a second obtaining module, a third obtaining module, an association module and an enhancement module, wherein:

[0238] The first obtaining module is configured to obtain point cloud data and camera data of a detection region.

[0239] The second obtaining module is configured to detect the point cloud data to obtain a detection result of the point cloud data.

[0240] The third obtaining module is configured to detect the camera data to obtain a detection result of the camera data.

[0241] The association module is configured to associate and match the detection result of the point cloud data and the detection result of the camera data according to a pre-established mapping relationship.

[0242] The enhancement module is configured to enhance the detection result of the point cloud data according to the association and matching result, to obtain an enhanced detection result of the point cloud data.

[0243] The point cloud data detection result enhancement device provided in this embodiment can execute the above-mentioned method embodiments, and has similar implementation principles and technical effects, which will not be described here again.

[0244] On the basis of the above-mentioned embodiment, optionally, the association module comprises: a first acquisition unit, a first calculation unit and a first matching unit, wherein:

[0245] The first acquisition unit is configured to map the detection result of the point cloud data to the camera data according to the mapping relationship to obtain a mapping result of the detection result of the point cloud data.

[0246] The first calculation unit is configured to calculate a first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data.

[0247] The first matching unit is configured to perform associated matching on the detection result of the point cloud data and the detection result of the camera data according to the first similarity.

[0248] The point cloud data detection result enhancement device provided in the embodiment can execute the above-mentioned method embodiment, and has similar implementation principles and technical effects, which will not be described here.

[0249] On the basis of the above-mentioned embodiment, optionally, the association module comprises: a second acquisition unit, a second calculation unit and a second matching unit, wherein:

[0250] The second acquisition unit is configured to map the camera data into a point cloud data coordinate system according to the mapping relationship to obtain a mapping result of the camera data.

[0251] The second calculation unit is configured to calculate a second similarity between the mapping result of the camera data and the detection result of the point cloud data.

[0252] The second matching unit is configured to perform associated matching on the detection result of the point cloud data and the detection result of the camera data according to the second similarity.

[0253] The point cloud data detection result enhancement device provided in the embodiment can execute the above-mentioned method embodiment, and has similar implementation principles and technical effects, which will not be described here.

[0254] On the basis of the above-mentioned embodiment, optionally, the enhancement module comprises: an enhancement unit, wherein:

[0255] The enhancement unit is configured to determine the detection result of the camera data as an enhanced detection result of the point cloud data if the associated matching result is that the detection result of the point cloud data and the detection result of the camera data are not matched successfully.

[0256] The point cloud data detection result enhancement device provided in the embodiment can execute the above-mentioned method embodiment, and has similar implementation principles and technical effects, which will not be described here.

[0257] On the basis of the above-mentioned embodiment, optionally, the device further comprises a first determination module, a second determination module, a fourth acquisition module and a fifth acquisition module, wherein:

[0258] The first determination module is configured to determine a positioning point of the detection result of the camera data in the point cloud coordinate system according to the positioning frame of the detection result of the camera data.

[0259] The second determination module is configured to determine the position of the detection result of the camera data in the point cloud coordinate system according to the positioning point and the pre-established fitting positioning relationship.

[0260] The fourth acquisition module is configured to correlate and match the detection result of the camera data by using a video tracking algorithm and the position of the detection result of the camera data in the point cloud coordinate system, to obtain a driving track of the detection result of the camera data in the point cloud coordinate system.

[0261] The fifth acquisition module is configured to obtain feature information of the detection result of the camera data according to the driving track of the detection result of the camera data in the point cloud coordinate system; the feature information of the detection result of the camera data comprises at least one of the position of the detection result of the camera data, the speed of the detection result of the camera data, the heading angle of the detection result of the camera data and the acceleration of the detection result of the camera data.

[0262] The point cloud data detection result enhancement device provided in the embodiment can execute the method embodiments, and has similar implementation principles and technical effects, which will not be described here.

[0263] On the basis of the above-mentioned embodiment, optionally, the device further comprises a sixth acquisition module, wherein:

[0264] The sixth acquisition module is configured to obtain the size of the detection result of the camera data according to the size of the positioning frame of the detection result of the camera data.

[0265] The point cloud data detection result enhancement device provided in the embodiment can execute the method embodiments, and has similar implementation principles and technical effects, which will not be described here.

[0266] On the basis of the above-mentioned embodiment, optionally, the device further comprises a seventh acquisition module, an eighth acquisition module, a ninth acquisition module, a tenth acquisition module, a first establishment module, a fitting module, a matching module and a second establishment module, wherein:

[0267] The seventh acquisition module is configured to acquire second sample point cloud data.

[0268] The eighth obtaining module is configured to convert the second sample point cloud data to the camera coordinate system by using an external conversion matrix of the laser radar to the camera coordinate system, to obtain processed second sample point cloud data; the processed second sample point cloud data is second sample point cloud data from which points not within the camera collection angle and points not successfully received by the laser radar are filtered out.

[0269] The ninth obtaining module is configured to convert the processed second sample point cloud data to a polar coordinate system, to obtain a polar angle of the processed second sample point cloud data.

[0270] The tenth obtaining module is configured to add a filling point to two points with equal polar angles of the processed second sample point cloud data by equally dividing a polar length, to obtain filled second sample point cloud data.

[0271] The first establishing module is configured to convert the filled second sample point cloud data from a point cloud coordinate system to a pixel coordinate system, to establish a mapping relationship pair of a depth point of the filled second sample point cloud data and a pixel point in the pixel coordinate system.

[0272] The fitting module is configured to fit pixel coordinate information corresponding to the filled second sample point cloud data according to the mapping relationship pair of the depth point of the filled second sample point cloud data and the pixel point in the pixel coordinate system, to obtain a calibration parameter.

[0273] The matching module is configured to match third sample point cloud data and third sample camera data of a same sample detection region by using the calibration parameter.

[0274] The second establishing module is configured to establish a mapping relationship according to a matching result of the third sample point cloud data and the third sample camera data.

[0275] The point cloud data detection result enhancement device provided in this embodiment can execute the method embodiments, and has similar implementation principles and technical effects, which will not be described herein again.

[0276] The specific limitations of the point cloud data detection result enhancement device can be referred to the limitations of the point cloud data detection result enhancement method, which will not be described herein again. The modules in the point cloud data detection result enhancement device can be all or partially realized by software, hardware, and combinations thereof. The modules can be embedded in or independent of a processor in a computer device in a hardware form, or stored in a memory in a computer device in a software form, so as to be called and executed by a processor to perform operations corresponding to the modules.

[0277] In one embodiment, a computer device is provided, including a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the following steps:

[0278] Obtain point cloud data of a detection region;

[0279] input the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; the detection model is obtained by mapping first sample camera data into a point cloud coordinate system, correlatively matching a detection result of the first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to a correlatively matching result; the first sample camera data and the first sample point cloud data are data of a same detection region.

[0280] The computer device provided in the above embodiments has similar implementation principles and technical effects to the above method embodiments, and thus detailed description is omitted here.

[0281] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor implementing the following steps when executing the computer program:

[0282] obtaining point cloud data and camera data of a detection region;

[0283] detecting the point cloud data to obtain a detection result of the point cloud data;

[0284] detecting the camera data to obtain a detection result of the camera data;

[0285] correlatively matching the detection result of the point cloud data and the detection result of the camera data according to a pre-established mapping relationship;

[0286] enhancing the detection result of the point cloud data according to a correlatively matching result to obtain an enhanced detection result of the point cloud data.

[0287] The computer device provided in the above embodiments has similar implementation principles and technical effects to the above method embodiments, and thus detailed description is omitted here.

[0288] In one embodiment, a computer readable storage medium is provided, storing a computer program, and the computer program is executed by a processor to implement the following steps:

[0289] obtaining point cloud data of a detection region;

[0290] The point cloud data is input into a preset detection model to obtain an enhanced detection result of the point cloud data; the detection model is obtained by mapping first sample camera data into a point cloud coordinate system, correlatively matching a detection result of the first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data, and training a preset initial detection model according to a correlatively matching result; the first sample camera data and the first sample point cloud data are data of a same detection region.

[0291] The computer readable storage medium provided in the above embodiments has similar implementation principles and technical effects to the above method embodiments, and thus details are not described herein.

[0292] In one embodiment, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program. When the computer program is executed by a processor, the following steps are implemented:

[0293] Obtaining point cloud data and camera data of a detection region;

[0294] Detecting the point cloud data to obtain a detection result of the point cloud data;

[0295] Detecting the camera data to obtain a detection result of the camera data;

[0296] Correlatively matching the detection result of the point cloud data and the detection result of the camera data according to a preset mapping relationship;

[0297] Enhancing the detection result of the point cloud data according to a correlatively matching result to obtain an enhanced detection result of the point cloud data.

[0298] The computer readable storage medium provided in the above embodiments has similar implementation principles and technical effects to the above method embodiments, and thus details are not described herein.

[0299] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiment methods. Any reference to memory, storage, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0300] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, but as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0301] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent. It should be pointed out that for those skilled in the art, without departing from the concept of the present application, some modifications and improvements can be made, which are all within the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A method for enhancing a point cloud data detection result, characterized in that, The method comprises: Obtaining point cloud data of a detection area; Inputting the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; wherein the detection model is a model that maps first sample camera data into a point cloud coordinate system, correlates and matches a detection result of the first sample point cloud data and a detection result of the first sample camera data according to a similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data to obtain a correlation matching result, labels the first sample point cloud data according to the correlation matching result to obtain a label of the first sample point cloud data; and the initial detection model is trained according to the first sample point cloud data and the label of the first sample point cloud data to obtain the detection model; the first sample camera data and the first sample point cloud data are data of the same detection area.

2. The method of claim 1, wherein, The training method of the detection model comprises: Mapping the detection result of the first sample point cloud data onto the first sample camera data according to a pre-established mapping relationship to obtain a mapping result of the detection result of the first sample point cloud data; Calculating a first similarity between the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data; Correlating and matching the mapping result of the detection result of the first sample point cloud data and the detection result of the first sample camera data according to the first similarity; Labeling the first sample point cloud data according to the correlation matching result to obtain a label of the first sample point cloud data; Training the initial detection model according to the first sample point cloud data and the label of the first sample point cloud data to obtain the detection model.

3. The method of claim 1, wherein, The training method of the detection model comprises: Mapping the first sample camera data into a point cloud data coordinate system according to a pre-established mapping relationship to obtain a mapping result of the first sample camera data; Calculating a second similarity between the mapping result of the first sample camera data and the detection result of the first sample point cloud data; Correlating and matching the detection result of the first sample point cloud data and the mapping result of the first sample camera data according to the second similarity; Labeling the first sample point cloud data according to the correlation matching result to obtain a label of the first sample point cloud data; Training the initial detection model according to the first sample point cloud data and the label of the first sample point cloud data to obtain the detection model.

4. The method according to claim 2 or 3, characterized in that, The method further comprises: Obtaining second sample point cloud data; Converting the second sample point cloud data into a camera coordinate system through an external conversion matrix of the laser radar to camera coordinate system to obtain processed second sample point cloud data; the processed second sample point cloud data is point cloud data filtered out of points not within the camera capture angle and points not successfully received by the laser radar; Converting the processed second sample point cloud data into a polar coordinate system to obtain a polar angle of the processed second sample point cloud data; The two points with equal polar angles of the processed second sample point cloud data are added with padding points by bisecting the polar length to obtain padded second sample point cloud data; The padded second sample point cloud data is converted from a point cloud coordinate system to a pixel coordinate system to establish a mapping relationship pair between depth points of the padded second sample point cloud data and pixel points in the pixel coordinate system; According to the mapping relationship pair between the depth points of the padded second sample point cloud data and the pixel points in the pixel coordinate system, pixel coordinate information corresponding to the padded second sample point cloud data is fitted to obtain calibration parameters; The third sample point cloud data and the third sample camera data of the same sample detection region are matched by using the calibration parameters; The mapping relationship is established according to the matching result of the third sample point cloud data and the third sample camera data.

5. A method for enhancing a point cloud data detection result, characterized in that, The method comprises: Obtaining point cloud data and camera data of a detection region; Detecting the point cloud data to obtain a detection result of the point cloud data; Detecting the camera data to obtain a detection result of the camera data; According to a pre-established mapping relationship, the detection result of the point cloud data and the detection result of the camera data are associated and matched to obtain an associated matching result, and if the associated matching result is that the detection result of the point cloud data and the detection result of the camera data are not successfully matched, the detection result of the camera data is determined as an enhanced detection result of the point cloud data.

6. The method of claim 5, wherein, According to the mapping relationship, the detection result of the point cloud data is mapped to the camera data to obtain a mapping result of the detection result of the point cloud data; A first similarity between the mapping result of the detection result of the point cloud data and the detection result of the camera data is calculated; According to the first similarity, the detection result of the point cloud data and the detection result of the camera data are associated and matched. According to the mapping relationship, the camera data is mapped to a point cloud data coordinate system to obtain a mapping result of the camera data; 7. The method of claim 5, wherein, A second similarity between the mapping result of the camera data and the detection result of the point cloud data is calculated; According to the second similarity, the detection result of the point cloud data and the detection result of the camera data are associated and matched. The method further comprises: According to a positioning frame of the detection result of the camera data, a positioning point of the detection result of the camera data in a point cloud coordinate system is determined; 8. The method according to claim 6 or 7, characterized in that, According to the positioning point and a pre-established fitting positioning relationship, a position of the detection result of the camera data in the point cloud coordinate system is determined; Using a video tracking algorithm and the position of the detection result of the camera data in the point cloud coordinate system, the detection result of the camera data is associated and matched to obtain a driving trajectory of the detection result of the camera data in the point cloud coordinate system; ​ ​ According to the detection result of the camera data, a driving track in the point cloud coordinate system is obtained, and feature information of the detection result of the camera data is obtained; the feature information of the detection result of the camera data includes at least one of position, speed, heading angle and acceleration of the detection result of the camera data.

9. The method of claim 8, wherein, The method further includes: According to the size of the bounding box of the detection result of the camera data, the size of the detection result of the camera data is obtained.

10. The method of claim 6 or 7, wherein, The method further includes: Obtaining second sample point cloud data; The second sample point cloud data is converted to a camera coordinate system through an external conversion matrix of the laser radar to the camera coordinate system, and processed second sample point cloud data is obtained; the processed second sample point cloud data is filtered out of points not within the camera's collection angle and points not successfully received by the laser radar; The processed second sample point cloud data is converted to a polar coordinate system to obtain the polar angle of the processed second sample point cloud data; Two points with equal polar angles of the processed second sample point cloud data are added with filling points by bisecting the polar length to obtain filled second sample point cloud data; The filled second sample point cloud data is converted from the point cloud coordinate system to the pixel coordinate system to establish a mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system; According to the mapping relationship pair between the depth points of the filled second sample point cloud data and the pixel points in the pixel coordinate system, the pixel coordinate information corresponding to the filled second sample point cloud data is fitted to obtain calibration parameters; Using the calibration parameters, third sample point cloud data and third sample camera data of the same sample detection region are matched; According to the matching results of the third sample point cloud data and the third sample camera data, the mapping relationship is established.

11. An enhancement device for point cloud data detection results, characterized in that, The device includes: An acquisition module is configured to acquire point cloud data of a detection region; An enhancement module is configured to input the point cloud data into a preset detection model to obtain an enhanced detection result of the point cloud data; wherein the detection model is configured to map first sample camera data to a point cloud coordinate system, correlate and match a detection result of the first sample point cloud data and a detection result of the first sample camera data according to a similarity between a mapping result of the first sample camera data and a mapping result of the first sample point cloud data to obtain a correlation and matching result, label the first sample point cloud data according to the correlation and matching result to obtain a label of the first sample point cloud data; and the first sample point cloud data and the first sample camera data are data of the same detection region.

12. An apparatus for enhancing a point cloud data detection result, the apparatus comprising: a point cloud data detection result enhancement module configured to enhance a point cloud data detection result. The device includes: A first acquisition module is configured to acquire point cloud data and camera data of a detection region; A second acquisition module is configured to detect the point cloud data to obtain a detection result of the point cloud data; A third obtaining module, detecting the camera data to obtain a detection result of the camera data; An association module, according to a pre-established mapping relationship, performing association matching on the detection result of the point cloud data and the detection result of the camera data to obtain an association matching result; An enhancement module, configured to, if the association matching result is that the detection result of the point cloud data and the detection result of the camera data are not successfully matched, determining the detection result of the camera data as an enhanced detection result of the point cloud data.

13. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method in any one of claims 1 to 10.

14. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 10.

Images