Vehicle detection method, device and system and computer readable storage medium

Abstract

The invention relates to a vehicle detection method, device and system, and a computer readable storage medium, and relates to the field of vehicle security check. The vehicle detection method comprises the following steps: in a process that at least one to-be-detected vehicle runs on a detection road section, acquiring original point cloud data by using a multi-line laser radar, and determining a laser scanning line to which each point cloud of each to-be-detected vehicle in the original point cloud data belongs; according to the original point cloud data, obtaining vehicle point cloud data of each to-be-detected vehicle in the at least one to-be-detected vehicle; according to the vehicle point cloud data, determining at least one of position information and vehicle speed information of each to-be-detected vehicle; determining target point cloud data of each to-be-detected vehicle obtained by scanning at least one laser line of which the vertical field angle is greater than a vertical field angle threshold value from the vehicle point cloud data; and according to the target point cloud data of each to-be-detected vehicle, determining position information of a loaded object of each to-be-detected vehicle.

Description

Technical Field

[0001] The present disclosure relates to the field of vehicle security inspection, and in particular to a vehicle inspection method, device and system, and a computer-readable storage medium. Background Art

[0002] When the vehicle safety inspection equipment is working, the position and speed information of the vehicle to be inspected needs to be captured in real time by sensors when the vehicle passes through the scanning channel. For container trucks, the position of the container on the vehicle also needs to be monitored simultaneously. These key position information will be transmitted to the scanning inspection system to provide accurate triggering timing for peripheral devices that perform X-ray scanning inspection processes, license plate recognition, container number recognition, etc.

[0003] In the related technology, a single-line laser radar is installed near the height of the vehicle license plate to detect the front and rear positions of the vehicle, and another single-line laser radar is installed at a height position where the container body can be stably observed to detect the front and rear surfaces of the container. Summary of the invention

[0004] In the related art, when detecting multiple types of vehicles in the driving direction, more than two single-line laser radars are required, which increases the complexity of the system and increases the system cost.

[0005] In response to the above technical problems, the present disclosure proposes a solution that can reduce the complexity and cost of vehicle detection, improve the flexibility and accuracy of detection, and achieve unified detection of different vehicle models.

[0006] According to a first aspect of the present disclosure, a vehicle detection method is provided, comprising:

[0007] When at least one vehicle to be inspected is traveling on the inspection section, original point cloud data is acquired by using a multi-line laser radar, and a laser scanning line to which each point cloud of each vehicle to be inspected in the original point cloud data belongs is determined, wherein the multi-line laser radar is installed on the side of the inspection section for scanning the at least one vehicle to be inspected;

[0008] Acquire vehicle point cloud data of each of the at least one vehicle to be inspected according to the original point cloud data;

[0009] Determine at least one of the position information and the speed information of each vehicle to be inspected according to the vehicle point cloud data;

[0010] Determine, from the vehicle point cloud data, target point cloud data of each vehicle to be inspected obtained by scanning with at least one laser line having a vertical field of view angle greater than a vertical field of view angle threshold;

[0011] The position information of the load of each vehicle to be inspected is determined according to the target point cloud data of each vehicle to be inspected.

[0012] According to a second aspect of the present disclosure, a vehicle detection device is provided, comprising:

[0013] A first acquisition module is configured to acquire original point cloud data using a multi-line laser radar when at least one vehicle to be inspected is traveling on a detection section, and determine a laser scanning line to which each point cloud of each vehicle to be inspected in the original point cloud data belongs, wherein the multi-line laser radar is installed on the side of the detection section for scanning the at least one vehicle to be inspected;

[0014] A second acquisition module is configured to acquire vehicle point cloud data of each of the at least one vehicle to be inspected according to the original point cloud data;

[0015] A first determination module is configured to determine at least one of the position information and the speed information of each vehicle to be inspected according to the vehicle point cloud data;

[0016] A second determination module is configured to determine, from the vehicle point cloud data, target point cloud data of each vehicle to be inspected obtained by scanning with at least one laser line having a vertical field of view angle greater than a vertical field of view angle threshold;

[0017] The third determination module is configured to determine the position information of the load of each vehicle to be inspected according to the target point cloud data of each vehicle to be inspected.

[0018] According to a third aspect of the present disclosure, a vehicle detection device is provided, comprising: a memory; and a processor coupled to the memory, wherein the processor is configured to execute the vehicle detection method described in any of the above embodiments based on instructions stored in the memory.

[0019] According to a fourth aspect of the present disclosure, a vehicle detection system is provided, comprising:

[0020] The vehicle detection device according to any one of the above embodiments;

[0021] The multi-line laser radar is configured to scan at least one vehicle to be inspected traveling on a detection section, obtain original point cloud data of the at least one vehicle to be inspected, and send the original point cloud data to the vehicle detection device.

[0022] According to a fifth aspect of the present disclosure, a computer-readable storage medium is provided, on which computer program instructions are stored, and when the instructions are executed by a processor, the vehicle detection method described in any of the above embodiments is implemented.

[0023] According to the present disclosure, the complexity and cost of vehicle detection can be reduced, the flexibility and accuracy of detection can be improved, and unified detection of different vehicle models can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings, which constitute a part of the specification, illustrate embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0025] The present disclosure may be more clearly understood from the following detailed description with reference to the accompanying drawings, in which:

[0026] Figure 1 is a schematic diagram showing a flow chart of a vehicle detection method according to some embodiments of the present disclosure;

[0027] Figure 2 is a schematic diagram showing a process of obtaining vehicle point cloud data of each vehicle to be inspected according to some embodiments of the present disclosure;

[0028] Figure 3 is a schematic diagram showing a process of determining the position information of the load of each vehicle to be inspected according to some embodiments of the present disclosure;

[0029] Figure 4 is a schematic diagram showing a multi-line laser radar according to some embodiments of the present disclosure;

[0030] Figure 5 is a schematic diagram showing a multi-line laser radar scanning a vehicle to be inspected according to some embodiments of the present disclosure;

[0031] Figure 6 is a schematic diagram of point cloud data showing the determination of the attribution of each point cloud according to some embodiments of the present disclosure;

[0032] Figure 7 is a schematic diagram showing point cloud data of a vehicle to be inspected according to some embodiments of the present disclosure;

[0033] Figure 8 is a schematic diagram showing an obliquely installed multi-line laser radar according to some embodiments of the present disclosure;

[0034] Fig. 9 is a schematic diagram showing a horizontally mounted multi-line laser radar according to some embodiments of the present disclosure;

[0035] Fig.10 is a schematic diagram showing original point cloud data before correction according to some embodiments of the present disclosure;

[0036] Fig.11 is a schematic diagram showing corrected point cloud data after correction according to some embodiments of the present disclosure;

[0037] Fig.12is a diagram showing the effect of multi-vehicle segmentation according to some embodiments of the present disclosure;

[0038] Fig.13 is a schematic flow chart showing a vehicle detection method according to other embodiments of the present disclosure;

[0039] Fig.14 is a block diagram showing a vehicle detection device according to some embodiments of the present disclosure;

[0040] Fig.15 is a block diagram showing a vehicle detection device according to some other embodiments of the present disclosure;

[0041] Fig.16 is a block diagram illustrating a vehicle detection system according to some embodiments of the present disclosure;

[0042] Fig.17 is a block diagram illustrating a computer system for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0043] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present disclosure unless otherwise specifically stated.

[0044] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0045] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present disclosure, its application, or uses.

[0046] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered as part of the specification.

[0047] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0048] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0049] Figure 1 It is a flowchart showing a vehicle detection method according to some embodiments of the present disclosure.

[0050] like Figure 1As shown, the vehicle detection method includes: step S110, in the process of at least one vehicle to be inspected driving on the detection section, using a multi-line laser radar to obtain original point cloud data, and determine the laser scanning line to which each point cloud of each vehicle to be inspected in the original point cloud data belongs, wherein the multi-line laser radar is installed on the side of the detection section for scanning at least one vehicle to be inspected; step S120, according to the original point cloud data, obtaining vehicle point cloud data of each vehicle to be inspected in at least one vehicle to be inspected; step S130, according to the vehicle point cloud data, determining at least one of the position information and the speed information of each vehicle to be inspected; step S140, from the vehicle point cloud data, determining the target point cloud data of each vehicle to be inspected obtained by scanning at least one laser line with a vertical field of view angle greater than a vertical field of view angle threshold; and step S150, according to the target point cloud data of each vehicle to be inspected, determining the position information of the load of each vehicle to be inspected.

[0051] In this embodiment, since the multi-line laser radar has multiple laser lines in the vertical direction, the multi-line laser radar has a wider detection range in the vertical direction. Regardless of whether at least one vehicle to be inspected traveling on the detection section is of the same model or of a different model, the multi-line laser radar can detect the point cloud data of any vehicle to be inspected, and then determine at least one of the position information and speed information of any vehicle to be inspected, which can reduce the complexity and cost of vehicle detection, improve the flexibility and accuracy of detection, and achieve unified detection of different vehicles. In this disclosure, laser lines and laser scanning lines have the same meaning.

[0052] In addition, since the load of the vehicle to be inspected is usually high, the target point cloud data of each vehicle to be inspected obtained by scanning at least one laser line with a vertical field angle greater than the vertical field angle threshold can be determined from the vehicle point cloud data. Among them, at least one laser line comes from a multi-line laser radar. It can be understood that by limiting the vertical field angle, the target point cloud data obtained is the point cloud data associated with the load of the vehicle to be inspected. Furthermore, according to the target point cloud data of each vehicle to be inspected, the position information of the load of each vehicle to be inspected can be determined, which can reduce the processing amount of point cloud data when determining the position of the load, thereby reducing the complexity and cost of vehicle load detection as a whole, and improving the detection efficiency of the load. In addition, the target point cloud data is obtained by scanning at least one laser line with a vertical field angle greater than the vertical field angle threshold, which can reduce the interference caused by the point cloud data obtained by scanning the laser line less than the vertical field angle threshold in the vehicle point cloud data, and improve the detection accuracy of the load.

[0053] In the present disclosure, by only using a multi-line laser radar, the position and / or speed of the vehicle and the position of the load can be determined simultaneously based on the acquisition of original point cloud data, thereby further improving the detection efficiency of the vehicle and the load of the vehicle, improving the flexibility of detection, and reducing the detection complexity and cost of the vehicle and the load of the vehicle.

[0054] On the other hand, a single-line laser radar has only one laser line in the vertical direction, while a multi-line laser radar has multiple laser lines in the vertical direction, which significantly expands the vertical monitoring range of the multi-line laser radar. Compared with the single-line laser solution, the blind area is significantly reduced, which improves the accuracy of detection, reduces the stringent requirements on the radar installation position and tilt angle, and improves the flexibility and adaptability of radar installation.

[0055] In step S110, when at least one vehicle to be inspected is traveling on the detection section, the original point cloud data is obtained by using a multi-line laser radar, and the laser scanning line to which each point cloud in the original point cloud data belongs is determined, wherein the multi-line laser radar is installed on the side of the detection section for scanning at least one vehicle to be inspected. For example, the installation position of the multi-line laser radar ensures that the vehicle to be inspected is scanned when it is traveling on the detection section. Since the multi-line laser radar is installed on the side of the detection section, when the vehicle to be inspected enters the detection section, the multi-line laser radar can scan the vehicle to be inspected. As the vehicle to be inspected travels along the detection section, the multi-line laser radar continuously scans the vehicle to be inspected to obtain the original point cloud data of the vehicle to be inspected. It should be noted that the original point cloud data obtained by using the multi-line laser radar usually includes environmental point cloud data. When no vehicle to be inspected passes by, the original point cloud data only includes environmental point cloud data.

[0056] In some embodiments, the laser scanning line to which each point cloud of each vehicle to be inspected in the original point cloud data belongs may be determined in the following manner.

[0057] First, for each point cloud of each vehicle to be inspected in the original point cloud data, determine the ratio of the coordinate value of each point cloud on the vertical axis in the radar coordinate system with the multi-line laser radar as the origin to the projection length of the line segment formed by each point cloud and the origin on the plane formed by the first horizontal axis and the second horizontal axis in the radar coordinate system.

[0058] For example, the coordinate value of each point cloud on the vertical axis in the radar coordinate system with the multi-line laser radar as the origin represents the vertical distance of each point cloud to the plane formed by the first horizontal axis and the second horizontal axis.

[0059] For example, the vertical axis is the z axis, the first horizontal axis is the x axis, and the second horizontal axis is the y axis, wherein the z axis is perpendicular to the rotation axis of the multi-line laser radar, and the x axis and the y axis are parallel to the rotation axis of the multi-line laser radar.

[0060] Then, the difference between the ratio and the tangent of the vertical field angle of each of the at least one laser line is determined.

[0061] Finally, when the difference between the ratio of a point cloud and the tangent value of one of the at least one laser line is within the difference range, the point cloud is determined to be point cloud data belonging to the one laser line. For example, the difference range is -0.01 to 0.01. The difference range here is only used as an example and does not constitute a specific limitation to the present disclosure.

[0062] For each point cloud, when there is a difference value within the difference range among the differences corresponding to each point cloud, the ratio can be regarded as approximately equal to the tangent value of the vertical field of view corresponding to the difference value within the difference range, that is, each point cloud belongs to the point cloud in the original point cloud data obtained by scanning at least one laser line. In this case, the angle between the line segment formed by each point cloud and the origin and the plane formed by the first horizontal axis and the second horizontal axis is approximately equal to or equal to the vertical field of view angle of one of the at least one laser line. That is, the point cloud can be regarded as originating from the one laser line.

[0063] For each point cloud, when there is no difference within the difference range, it is determined that each point cloud does not belong to a point cloud in the original point cloud data obtained by the corresponding laser line scanning.

[0064] Through the above operations, the laser scanning line to which each point cloud in the original point cloud data belongs can be calculated.

[0065] In step S120, vehicle point cloud data of each of the at least one vehicle to be inspected is obtained based on the original point cloud data. For example, the original point cloud data may be processed by ground correction, conditional Euclidean clustering, and other operations described later to obtain vehicle point cloud data of each vehicle to be inspected.

[0066] For example, in some embodiments, Figure 2 Steps S121 to S124 shown are used to obtain the vehicle point cloud data of each of the at least one vehicle to be inspected based on the original point cloud data. Figure 2 It is a schematic diagram showing a process of obtaining vehicle point cloud data of each vehicle to be inspected according to some embodiments of the present disclosure.

[0067] In step S121, ground point cloud data is obtained from the original point cloud data, and a ground plane model of the detection section is constructed using a random sampling consistency algorithm. In step S122, a rotation matrix is ​​determined based on the normal vector of the ground plane model and the unit direction vector of the vertical axis of the radar coordinate system, wherein the rotation matrix rotates the normal vector to the vertical axis. In step S123, the original point cloud data is ground corrected using the rotation matrix to obtain corrected point cloud data. In step S124, the point cloud data in the channel is obtained based on the corrected point cloud data, and the point cloud data in the channel is clustered using the conditional Euclidean clustering algorithm to obtain the vehicle point cloud data of each vehicle to be inspected.

[0068] In some embodiments, when the point cloud data in the channel is clustered to obtain a non-vehicle object, the duration of the non-vehicle object's existence is monitored; when the duration is greater than a time threshold, an early warning is issued. For example, when the vehicle point cloud data is clustered to obtain a non-vehicle object, first detect whether the non-vehicle object is a person (i.e., human intrusion detection). If the non-vehicle object is a person, determine whether the non-vehicle object continues to exist. If the non-vehicle object continues to exist, a human intrusion alarm is issued. If the non-vehicle object does not continue to exist, the duration of the non-vehicle object's existence is recorded.

[0069] In step S130, at least one of the position information and the speed information of each vehicle to be inspected is determined based on the vehicle point cloud data. For example, the position and speed of the vehicle can be determined based on the analysis results by analyzing the position and timestamp changes of the point cloud between consecutive frames in the vehicle point cloud data of each vehicle to be inspected.

[0070] In step S140, target point cloud data of each vehicle to be inspected obtained by scanning at least one laser line with a vertical field of view angle greater than a vertical field of view angle threshold is determined from the vehicle point cloud data. The at least one laser line comes from a multi-line laser radar. For example, the at least one laser line can be a laser line with the largest vertical field of view angle, or a laser line with the largest vertical field of view angle (i.e., the laser line at the top layer of the scanning layer) and a laser line with the second largest vertical field of view angle (i.e., the laser line at the top layer of the scanning layer).

[0071] It should be noted that when the vehicle load is a container, the height range of the container is basically determined. Therefore, for a multi-line laser radar installed at a fixed height, its vertical field of view angle threshold is also basically determined, thereby obtaining point cloud data associated with the edge of the container (such as the front edge, rear edge, etc.). In this way, point cloud data that can stably observe a specific range of the container can be selected for analysis.

[0072] For example, when the vertical field of view angle threshold is 13 degrees and the maximum vertical field of view angle of the multi-line laser radar is 15 degrees, at least one laser line is a laser line with a vertical field of view angle between 13 degrees and 15 degrees.

[0073] In step S150, the position information of the load of each vehicle to be inspected is determined based on the target point cloud data of each vehicle to be inspected. For example, the load includes a load of a box structure, and may also be a load of other structures. For example, the load of a box structure may be a container, and may also be an object of other box structures. For example, the position information of the load may be determined by determining the front box surface and / or the rear box surface of the load of the box structure.

[0074] In some embodiments, the Figure 3 The method shown is used to determine the position information of the load of each vehicle to be inspected based on the target point cloud data of each vehicle to be inspected.

[0075] Figure 3 It is a schematic diagram showing a process of determining the location information of the load of each vehicle to be inspected according to some embodiments of the present disclosure.

[0076] like Figure 3 As shown, determining the position information of the load of each vehicle to be inspected includes steps S151 to S152.

[0077] In step S151, based on the target point cloud data of each vehicle to be inspected, it is determined whether there are discontinuous features and / or corner features in the target point cloud data of each vehicle to be inspected. A discontinuous feature refers to a situation where the distance between two discontinuous points in the target point cloud data is greater than a distance threshold. A corner feature refers to a situation where the curvature between points in the target point cloud data changes suddenly or dramatically.

[0078] For example, at least one of the discontinuity feature and the corner feature in the target point cloud data can be judged by feature analysis, feature recognition, etc. Specifically, the corner feature can be judged based on geometric analysis, such as by analyzing the curvature and normal vector between points. The position where these parameters change drastically may be a corner, and the neighborhood of the point that may be a corner is analyzed at the same time. For another example, discontinuity feature extraction can be performed by clustering algorithms, etc. to determine whether there is a discontinuity feature. For another example, corner feature detection can be performed by training models, deep learning, etc.

[0079] In step S152, when there are discontinuous features and / or corner features in the target point cloud data of each vehicle to be inspected, the position information of the load of each vehicle to be inspected is determined according to the point cloud coordinates of the points where the discontinuous features and / or corner features exist.

[0080] In some embodiments, the vehicle detection method may further include determining at least one of the front position, rear position and speed of the vehicle to be inspected based on the vehicle point cloud data. Since the multi-line laser radar has a wide vertical detection range, it can adapt to the detection of vehicles of different models and sizes, has strong vehicle compatibility, and the detection method is more versatile. In the present disclosure, by using only one multi-line laser radar, it is possible to accurately measure the position and speed of vehicles of various models, while effectively monitoring the position of vehicle loads (such as on-board containers) at different heights, ensuring that during the vehicle scanning process, various sensors such as area array cameras and line array cameras can efficiently and accurately capture and process vehicle data based on the determined position and speed of the vehicle and the position of the vehicle load. For example, various sensors such as area array cameras and line array cameras can select a suitable triggering time based on the determined position and speed of the vehicle and the position of the vehicle load.

[0081] The following will be combined Figures 4 to 14 The vehicle detection method in some embodiments of the present disclosure is described.

[0082] Figure 4 is a schematic diagram showing a multi-line laser radar according to some embodiments of the present disclosure.

[0083] For example, when the detection section is set as the scanning channel of the vehicle safety inspection equipment, similar Figure 4 The multi-line laser radar shown is arranged on the side wall of the scanning channel. When the vehicle to be inspected passes through the scanning channel, all scanning lines (also called laser lines) of the multi-line laser radar can generate a vehicle contour point cloud. Figure 4 The multi-line laser radar in is a 16-line laser radar, and its vertical detection field angle range is ±15°, and the vertical angular resolution is 2°. However, the present disclosure may also use other models of multi-line laser radars with different line numbers, vertical field ranges, and vertical angular resolutions. The deployment of a multi-line laser radar at the side wall of the scanning channel is only an embodiment of the present disclosure and does not constitute a limitation on the layout scenario of the present disclosure. For example, it is also possible to install a column on the side of the road, and then install a multi-line laser radar on the column.

[0084] In some embodiments, these vehicle contour point clouds can be used for detecting the front position and rear position of the vehicle to be inspected and calculating the vehicle speed.

[0085] In some embodiments, taking the container as an example, the topmost scanning line, i.e., the +15 degree scanning line, can be extracted from the scanning layer of the multi-line laser radar. The detection height of this scanning line changes with the change of the distance between the multi-line laser radar and the vehicle to be inspected. When the vehicle to be inspected first enters the scanning channel, it is far away from the multi-line laser radar. The detection height of the +15 degree scanning line of the multi-line laser radar is high, so that the multi-line laser radar can initially obtain the location information of the container at a higher height. As the vehicle to be inspected continues to approach the multi-line laser radar, the detection height of the +15 degree scanning line of the multi-line laser radar decreases, so that the multi-line laser radar can also obtain the location information of the container at a lower height. That is, the multi-line laser radar can obtain the location information of the front and rear surfaces of the vehicle-mounted container at different heights. The location information here can be determined according to the position of the intersection of the front (rear) surface of the container and the side facing the multi-line laser radar. Through the hierarchical analysis of the point cloud data, the location information of the vehicle and the location information of the vehicle-mounted container can also be obtained at the same time. The hierarchical analysis here refers to determining the location information of the on-vehicle container using the target point cloud data obtained from the top-level scanning line in the scanning layer, and determining the location information of the vehicle using the vehicle point cloud data of all scanning layers obtained by multi-line laser radar scanning.

[0086] Figure 5 is a schematic diagram showing a multi-line laser radar scanning a vehicle to be inspected according to some embodiments of the present disclosure.

[0087] like Figure 5 As shown, the vehicle to be inspected enters the scanning channel, and the multi-line laser radar emits laser lines from the side of the scanning channel (that is, the side of the vehicle to be inspected) to the vehicle to be inspected, thereby scanning the vehicle to be inspected. For the +15 degree laser line, refer to Figure 5 It can be found that the laser line can scan the container loaded on the vehicle to be inspected as the vehicle to be inspected moves in the scanning channel.

[0088] For example, from the original point cloud data of the vehicle to be inspected obtained by multi-line laser radar scanning, the method of determining the laser scanning line to which each point cloud of each vehicle to be inspected in the original point cloud data belongs can refer to Figure 6 . Figure 6 is a schematic diagram showing how to determine the attribution of each point cloud according to some embodiments of the present disclosure. It is noted that by determining the laser scanning line to which each point cloud belongs, it is convenient to extract the target point cloud data of each vehicle to be inspected in the subsequent process.

[0089] like Figure 6 As shown in the figure, the radar coordinate system with the multi-line laser radar as the origin includes the x-axis, y-axis and z-axis. i ,y i ,z i ) point cloud, if we calculate Then the point cloud can be regarded as the points obtained by scanning the +15 degree laser line. By traversing the point cloud data and performing the above calculations, all the points on the +15 degree laser line can be extracted, that is, Figure 6 For each other point cloud, the same processing method can be used to determine the laser scanning line to which each point cloud belongs. Thus, the laser scanning line to which all point cloud data belongs can be determined.

[0090] z i The point cloud coordinates are (x i ,y i, z i ) is the coordinate value of the point on the z-axis, The point cloud coordinates are (x i ,y i, z i ) and the origin on the plane formed by the x-axis and y-axis.

[0091] Figure 7 is a schematic diagram showing point cloud data of a vehicle to be inspected according to some embodiments of the present disclosure. Figure 7 As shown, the vehicle point cloud data of the vehicle to be inspected includes the point cloud data of the vehicle head, tires, container, and the connecting parts between the vehicle head and the container. The contour features of the vehicle-mounted container on the scanning line include obvious corners (corner features) and / or the curved gap between the front container surface and the vehicle head (discontinuous features). These features provide key information for accurately locating the position of the container surface. By comprehensively analyzing these features, the position of the container can be accurately determined.

[0092] In some embodiments, in actual product applications, due to limitations on the scanning channel width and / or vehicle type detection requirements, multi-line laser radars are installed in two ways: horizontally and tilted, with the latter being more common. When a multi-line laser radar is installed tilted, the point cloud coordinate system inside the multi-line laser radar is not parallel to the actual ground, which increases the complexity of point cloud data processing.

[0093] Figure 8 is a schematic diagram showing an obliquely mounted multi-line laser radar according to some embodiments of the present disclosure.

[0094] Fig. 9 is a schematic diagram showing a horizontally mounted multi-line laser radar according to some embodiments of the present disclosure.

[0095] In order to solve the technical problem that the complexity of point cloud data processing increases due to tilted installation, the present disclosure proposes to perform ground correction on the original point cloud data collected by the multi-line laser radar before using the point cloud data of the vehicle to be inspected.

[0096] For example, the RANSAC (Random Sample Consensus) algorithm can be used to accurately fit the ground point cloud. The random sampling consensus algorithm builds a model by randomly selecting samples in the data set and iterates to find the model that best represents the entire data set.

[0097] The process of obtaining the ground data model using the RANSAC algorithm can be referred to as follows.

[0098] In the first step, point cloud filtering is used to filter out the point cloud data of irrelevant objects such as the environment, walls, and buildings in the original point cloud data, and only the point cloud data set containing the ground is selected.

[0099] In the second step, a group of points are randomly selected as initial inliers in the screened point cloud dataset, and an initial plane model is fitted based on these points.

[0100] The third step is to calculate the distances from other points in the filtered point cloud data set to the fitted plane model, and include points whose distances are less than a preset threshold into the internal point set.

[0101] In the fourth step, the second and third steps are repeated until a predetermined number of iterations is reached or the size of the inlier set reaches a certain threshold, and finally the best ground plane model is determined.

[0102] The ground plane model fitted by the RANSAC algorithm can be expressed as Ax+By+Cz+D=0, where the normal vector of the ground plane model is (A, B, C). In the point cloud coordinate system, the unit direction vector of the z-axis is (0, 0, 1). In order to rotate the point cloud coordinate system to align with the ground, the present disclosure calculates a rotation matrix that can rotate the ground normal vector to the z-axis direction. The Rodriguez rotation formula can be used to calculate this rotation matrix R. In practical applications, it is usually not necessary to manually calculate the rotation matrix, and most mathematical and scientific computing libraries provide implementations of the Rodriguez formula. After the point cloud data is rotated according to the rotation matrix, the ground point cloud data can be aligned with the radar coordinate system. As a result, the original point cloud data can be ground corrected to obtain the corrected point cloud data, which is convenient for subsequent filtering out of the point cloud data outside the scanning channel according to the scanning channel size.

[0103] Preferably, in the above-mentioned process of acquiring the ground plane model, the original point cloud data is acquired when there is no object (for example, no vehicle) in the scanning channel. At this time, the original point cloud data only includes the environmental point cloud data. The point cloud data obtained after filtering is empty, and then it is judged to be empty when the position information and / or speed information of each vehicle to be inspected is determined according to the vehicle point cloud data described later. This is because when the original point cloud data is obtained when there is an object in the channel, the object will block the ground, resulting in inaccurate ground detection. Preferably, the acquisition of the above-mentioned ground plane model is only performed once in the early configuration stage of the system. After obtaining a set of rotation matrix parameters, only the parameters need to be called up later. This is because if all point cloud data are ground corrected, the real-time execution of the vehicle detection method will be reduced.

[0104] Fig.10 is a schematic diagram showing original point cloud data before correction according to some embodiments of the present disclosure.

[0105] Fig.11 is a schematic diagram showing corrected point cloud data according to some embodiments of the present disclosure.

[0106] like Fig.10 As shown in , before correction, the horizontal plane of the radar coordinate system (the plane formed by the x-axis and the y-axis) is not parallel to the ground. Fig.11 As shown, after correction, the horizontal plane of the radar coordinate system is parallel to the ground.

[0107] In this embodiment, after the ground correction of the original point cloud data is completed, it is convenient to detect the position of the vehicle and the position of the vehicle's load (such as a container) based on the radar coordinate system.

[0108] In some embodiments, the corrected point cloud data can also be subjected to point cloud filtering to filter out the environmental point cloud. The aforementioned ground correction operation facilitates the point cloud filtering in this embodiment. Thus, for example, only the in-channel point cloud data of the scanning channel can be retained from the corrected point cloud data.

[0109] In some embodiments, when there are multiple vehicles to be inspected in the scanning channel, a conditional Euclidean clustering algorithm is used to accurately segment the point cloud data of different vehicles to be inspected, that is, conditional Euclidean clustering is performed on the vehicle point cloud data or the point cloud data in the channel after ground correction and point cloud filtering. The conditional Euclidean clustering algorithm is a distance-based clustering method that classifies data points by calculating the Euclidean distance between data points to ensure that similar data points are classified into the same cluster. In vehicle safety inspections, in order to ensure the integrity of vehicle data collection, a safe distance must be maintained between the front and rear vehicles. Therefore, in the conditional Euclidean clustering algorithm, the distance between data points in the direction of vehicle travel (i.e., the safe distance of the vehicle in the direction of travel) is set as a clustering constraint. When the distance between two points in the direction of vehicle travel exceeds a preset threshold, the algorithm automatically divides these data points into different clusters.

[0110] Fig.12 2 is a diagram showing the effect of multi-vehicle segmentation according to some embodiments of the present disclosure.

[0111] like Fig.12 As shown, through the conditional Euclidean clustering algorithm, the precise positions of the point cloud data of different vehicles to be inspected, that is, the point cloud data of vehicle to be inspected 1 and vehicle to be inspected 2, can be determined.

[0112] In some embodiments, the vehicle detection method includes at least one of point cloud downsampling, ground correction, pass-through filtering, conditional Euclidean clustering, point cloud classification, non-vehicle body target processing, vehicle body target processing and vehicle-mounted container position detection, and preferably includes all of the above steps.

[0113] In point cloud downsampling, after obtaining the original point cloud data, voxel filtering is first performed to reduce the number of laser points per unit volume, thereby improving the efficiency of subsequent data processing.

[0114] In ground correction, the pre-saved rotation matrix parameters are read to rotate the downsampled original point cloud data to align with the ground and the coordinate axis.

[0115] In the through-filtering, a through-filter is applied to filter and retain the point cloud data only in the detection channel (scanning channel), and obtain the point cloud data in the channel.

[0116] In conditional Euclidean clustering, a conditional Euclidean clustering algorithm is performed on the point cloud data in the channel, and points with similar distances are classified into one category to determine the point cloud distribution in the scanning channel.

[0117] In point cloud classification, the clustered point cloud data is classified according to the size, quantity and other characteristics of the point cloud data to distinguish vehicles or other target objects. Thus, vehicle point cloud data and non-vehicle point cloud data are obtained.

[0118] In the processing of non-vehicle targets, if the point cloud type is determined to be non-vehicle, the anti-intrusion detection process will be initiated. When the non-vehicle target is detected to continue to exist, an early warning will be issued.

[0119] In vehicle body target processing, if the point cloud type is determined to be a vehicle body, at least one of the vehicle's position information and speed information can be determined based on the vehicle point cloud data. For example, the vehicle point cloud data of each vehicle to be inspected is used to determine at least one of the position information and speed information of each vehicle to be inspected. Specifically, the position and timestamp changes of the point clouds between consecutive frames in the vehicle point cloud data of each vehicle to be inspected are analyzed, and the position and speed of the vehicle are determined based on the analysis results. The displacement of the vehicle can be determined, for example, by using at least one of the differential method of directly comparing point cloud data, the feature matching method, the iterative nearest point algorithm, the normal distribution transformation algorithm, and the method based on deep learning.

[0120] For example, in the position detection of the on-vehicle container, for the vehicle body point cloud, a +15-degree scan line can be extracted, and the contour change of the scan line can be detected from the rear to the front of the vehicle to identify the positions of the front and rear surfaces of the on-vehicle container, that is, to determine the target point cloud data from the vehicle point cloud data, and determine the position information of the on-vehicle container accordingly.

[0121] The following will be combined Fig.13 The specific implementation of the vehicle detection method in some embodiments of the present disclosure is described in detail.

[0122] Fig.13 It is a flowchart showing a vehicle detection method according to some other embodiments of the present disclosure.

[0123] like Fig.13 As shown, the vehicle detection method includes steps S210 to S224.

[0124] In step S210, the original point cloud data is obtained. For example, a multi-line laser radar installed on the side of the detection section (such as a scanning channel) is used to scan at least one vehicle to be inspected traveling on the detection section to obtain vehicle point cloud data of at least one vehicle to be inspected. It can be understood that in the case where there is no vehicle to be inspected, the original point cloud data only includes environmental point cloud data.

[0125] In step S211, a ground plane model of the detection section is constructed, that is, the ground parameters are fitted using a random sampling consistency algorithm to obtain a normal vector of the ground plane model as the ground parameter.

[0126] In step S212, a rotation matrix is ​​calculated based on the normal vector and the unit direction vector of the vertical axis of the radar coordinate system. The rotation matrix can be saved locally.

[0127] In step S213, the laser scanning line to which each point cloud in the original point cloud data belongs is determined. At this time, only the attribution of each point cloud is calculated and marked so as to be extracted after the vehicle point cloud data is obtained later.

[0128] In step S214, the vehicle point cloud data is downsampled to obtain downsampled point cloud data. There is no strict execution order for steps S211 to S212, steps S213, and S214, and they can also be performed simultaneously. Those skilled in the art can set it according to the situation.

[0129] In step S215, the rotation matrix calculated in step S212 (for example, reading the locally stored rotation matrix) is used to perform ground correction on the downsampled point cloud data to obtain corrected point cloud data so that the horizontal plane of the radar coordinate system is parallel to the ground.

[0130] In step S216, a point cloud filtering operation is performed on the corrected point cloud data to obtain filtered point cloud data. The point cloud filtering operation can filter the environmental point cloud and improve the accuracy of vehicle detection. In the case where the detection section is a scanning channel, the point cloud data in the scanning channel can be obtained through step S216.

[0131] In step S217, conditional Euclidean clustering is performed on the filtered point cloud data. Conditional Euclidean clustering clusters the filtered point cloud data using the safe distance between vehicles in the driving direction as a condition.

[0132] In step S218, the clustered point cloud data is classified to identify whether the clustered point cloud data belongs to a vehicle body.

[0133] When the point cloud data after clustering belongs to the vehicle body, steps S219 to S221 are executed.

[0134] In step S219, the location information and speed information of the vehicle to be inspected are determined based on the clustered point cloud data.

[0135] In step S220, target point cloud data is extracted from the clustered point cloud data. The target point cloud data is, for example, point cloud data of the vehicle to be inspected obtained by scanning at least one laser line in a multi-line laser radar whose vertical field of view angle is greater than a vertical field of view angle threshold.

[0136] In step S221 , the position information of the load of the vehicle to be inspected is determined according to the target point cloud data of the vehicle to be inspected.

[0137] When the point cloud data after clustering belongs to a non-vehicle body, steps S222 to S225 are executed.

[0138] In step S222, human intrusion detection is performed.

[0139] In step S223, it is determined whether the intruder target continues to exist.

[0140] In the case where the intruder target continues to exist, step S224 is executed, in which a human intrusion alarm is issued.

[0141] In the case where the intrusion target does not exist continuously, step S225 is executed. In step S225, the existence time of the target is recorded, and the human intrusion detection is continued.

[0142] Fig.13 The illustrated embodiment is only a detailed example and does not constitute a specific limitation to the present disclosure.

[0143] The technical solution disclosed in the present invention has high measurement accuracy, strong anti-interference ability, and is not easily affected by noise points or data missing, thereby ensuring the stability and reliability of the overall detection effect.

[0144] In the above embodiment, a multi-line laser radar is used to achieve accurate position and speed measurement of the vehicle target in the scanning channel, which can ensure efficient and accurate data collection of the vehicle in a dynamic environment. The multi-line laser radar is used to accurately locate the container on the vehicle, and the spatial position of the container can be accurately captured.

[0145] So far, the vehicle detection method in some embodiments of the present disclosure has been described.

[0146] It should be noted that in some of the embodiments described, a laser scanning radar is used to achieve accurate position and speed measurement of vehicles of various models, while effectively monitoring the position of the vehicle's load. However, in the case where, for example, the detection road (such as a scanning channel) is long, in order to increase the detection distance, more than one laser scanning radar may be used, such as two laser scanning radars arranged at intervals in the direction of vehicle travel. In this case, the vehicle detection method of the present disclosure can still be implemented, the difference being that before acquiring the vehicle point cloud data, the point cloud data of multiple multi-line laser radars are fused into a single point cloud data, and then the position information and / or vehicle speed information are confirmed, as well as the position information of the vehicle's load; or, after separately determining the vehicle's position information and / or vehicle speed information, and determining the position information of the vehicle's load, the relevant information is fused at the processing level.

[0147] Fig.14 is a block diagram illustrating a vehicle detection apparatus according to some embodiments of the present disclosure.

[0148] like Fig.14As shown, the vehicle detection device 13 includes a first acquisition module 131 , a second acquisition module 132 , a first determination module 133 , a second determination module 134 and a third determination module 135 .

[0149] The first acquisition module 131 is configured to acquire raw point cloud data using a multi-line laser radar when at least one vehicle to be inspected is traveling on the inspection section, and determine the laser scanning line to which each point cloud of each vehicle to be inspected in the raw point cloud data belongs. The multi-line laser radar is installed on the side of the inspection section to scan at least one vehicle to be inspected, for example, Figure 1 Step S110 is shown.

[0150] The second acquisition module 132 is configured to acquire the vehicle point cloud data of each of the at least one vehicle to be inspected according to the original point cloud data, for example, by performing the following steps: Figure 1 Step S120 is shown.

[0151] The first determination module 133 is configured to determine at least one of the position information and the speed information of each vehicle to be inspected according to the vehicle point cloud data, for example, by performing the following steps: Figure 1 Step S130 is shown.

[0152] The second determination module 134 is configured to determine, from the vehicle point cloud data, target point cloud data of each vehicle to be inspected obtained by scanning at least one laser line whose vertical field of view angle is greater than a vertical field of view angle threshold, for example, by performing the following steps: Figure 1 Step S140 is shown, wherein at least one laser line comes from a multi-line laser radar.

[0153] The third determination module 135 is configured to determine the position information of the load of each vehicle to be inspected according to the target point cloud data of each vehicle to be inspected, for example, by performing the following steps: Figure 1 Step S150 is shown.

[0154] Fig.15 is a block diagram showing a vehicle detection device according to some other embodiments of the present disclosure.

[0155] like Fig.15 As shown, the vehicle detection device 14 includes a memory 141; and a processor 142 coupled to the memory 141. The memory 141 is used to store instructions for executing the corresponding embodiments of the vehicle detection method. The processor 142 is configured to execute the vehicle detection method in any of the embodiments of the present disclosure based on the instructions stored in the memory 141.

[0156] Fig.16 is a block diagram illustrating a vehicle detection system according to some embodiments of the present disclosure.

[0157] like Fig.16As shown, the vehicle detection system 1 includes a vehicle detection device 13 and a multi-line laser radar 15 .

[0158] The vehicle detection device 13 is a vehicle detection device in any embodiment of the present disclosure, and is configured to execute the vehicle detection method in any embodiment of the present disclosure.

[0159] The multi-line laser radar 15 is configured to scan at least one vehicle to be inspected traveling on the inspection section, obtain the original point cloud data of at least one vehicle to be inspected, and send the original point cloud data to the vehicle detection device 13. The multi-line laser radar 15 is installed on the side of the inspection section.

[0160] Fig.17 is a block diagram illustrating a computer system for implementing some embodiments of the present disclosure.

[0161] like Fig.17 As shown, the computer system 160 may be embodied in the form of a general-purpose computing device. The computer system 160 includes a memory 1610, a processor 1620, and a bus 1600 that connects the various system components.

[0162] The memory 1610 may include, for example, a system memory, a non-volatile storage medium, etc. The system memory may store, for example, an operating system, an application program, a boot loader, and other programs. The system memory may include a volatile storage medium, such as a random access memory (RAM) and / or a cache memory. The non-volatile storage medium may store, for example, instructions for executing at least one of the corresponding embodiments of the vehicle detection method. The non-volatile storage medium includes, but is not limited to, a disk memory, an optical memory, a flash memory, etc.

[0163] The processor 1620 can be implemented by a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistors, etc. Discrete hardware components. Accordingly, each module such as the judgment module and the determination module can be implemented by a central processing unit (CPU) running instructions in a memory that execute corresponding steps, or can be implemented by a dedicated circuit that executes corresponding steps.

[0164] The bus 1600 may use any of a variety of bus architectures, including, but not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MCA) bus, and a Peripheral Component Interconnect (PCI) bus.

[0165] The computer system 160 may also include an input / output interface 1630, a network interface 1640, a storage interface 1650, etc. These interfaces 1630, 1640, 1650, the memory 1610, and the processor 1620 may be connected via a bus 1600. The input / output interface 1630 may provide a connection interface for input / output devices such as a display, a mouse, and a keyboard. The network interface 1640 may provide a connection interface for various networked devices. The storage interface 1650 may provide a connection interface for external storage devices such as a floppy disk, a USB flash drive, and an SD card.

[0166] Here, various aspects of the present disclosure are described with reference to flowcharts and / or block diagrams of methods, devices, and computer program products according to embodiments of the present disclosure. It should be understood that each frame of the flowchart and / or block diagram and the combination of frames can be implemented by computer-readable program instructions.

[0167] These computer-readable program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable device to produce a machine, so that the processor executes the instructions to produce means for implementing the functions specified in one or more blocks in the flowchart and / or block diagram.

[0168] These computer-readable program instructions may also be stored in a computer-readable memory, which cause the computer to work in a specific manner to produce an article of manufacture, including instructions for implementing the functions specified in one or more blocks in the flowchart and / or block diagram.

[0169] The present disclosure can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects.

[0170] The vehicle detection method, device and system, and computer-readable storage medium in the above-mentioned embodiments can reduce the complexity and cost of vehicle detection, improve the flexibility and accuracy of detection, and achieve unified detection of different vehicle models.

[0171] So far, the vehicle detection method, device and system, and computer-readable storage medium according to the present disclosure have been described in detail. In order to avoid obscuring the concept of the present disclosure, some details known in the art are not described. Based on the above description, those skilled in the art can fully understand how to implement the technical solution disclosed herein.

Claims

1. A vehicle detection method, comprising: When at least one vehicle to be inspected is traveling on the inspection section, original point cloud data is acquired by using a multi-line laser radar, and a laser scanning line to which each point cloud of each vehicle to be inspected in the original point cloud data belongs is determined, wherein the multi-line laser radar is installed on the side of the inspection section for scanning the at least one vehicle to be inspected; Acquire vehicle point cloud data of each of the at least one vehicle to be inspected according to the original point cloud data; Determine at least one of the position information and the speed information of each vehicle to be inspected according to the vehicle point cloud data; Determine, from the vehicle point cloud data, target point cloud data of each vehicle to be inspected obtained by scanning at least one laser line having a vertical field of view angle greater than a vertical field of view angle threshold; The position information of the load of each vehicle to be inspected is determined according to the target point cloud data of each vehicle to be inspected.

2. The vehicle detection method according to claim 1, wherein: Determining the position information of the load of each vehicle to be inspected according to the target point cloud data of each vehicle to be inspected includes: According to the target point cloud data of each vehicle to be inspected, determining whether there is a discontinuity feature and / or a corner feature in the target point cloud data of each vehicle to be inspected; When there are discontinuous features and / or corner features in the target point cloud data of each vehicle to be inspected, the position information of the load of each vehicle to be inspected is determined according to the point cloud coordinates of the points where the discontinuous features and / or corner features exist.

3. The vehicle detection method according to claim 1, wherein: Determining the laser scanning line to which each point cloud of each vehicle to be inspected in the original point cloud data belongs includes: For each point cloud of each vehicle to be inspected in the original point cloud data, determine a ratio of a coordinate value of each point cloud on a vertical axis in a radar coordinate system with the multi-line laser radar as the origin to a projection length of a line segment formed by each point cloud and the origin on a plane formed by a first horizontal axis in the radar coordinate system; determining a difference between the ratio and a tangent of a vertical field angle of each of the at least one laser line; When the difference between the ratio of a point cloud and the tangent value of the vertical field angle of one of the at least one laser line is within a difference range, it is determined that the point cloud belongs to the point cloud data emitted by the laser line.

4. The vehicle detection method according to claim 1, wherein: Acquiring the vehicle point cloud data of each of the at least one vehicle to be inspected according to the original point cloud data comprises: Obtaining ground point cloud data from the original point cloud data, and constructing a ground plane model of the detection section using a random sampling consistency algorithm; Determine a rotation matrix according to the normal vector of the ground plane model and the unit direction vector of the vertical axis of the radar coordinate system, wherein the rotation matrix rotates the normal vector to the vertical axis; Using the rotation matrix, ground correction is performed on the original point cloud data to obtain corrected point cloud data; According to the corrected point cloud data, the point cloud data in the channel is obtained, and the point cloud data in the channel is clustered using a conditional Euclidean clustering algorithm to obtain vehicle point cloud data of each vehicle to be inspected.

5. The vehicle detection method according to claim 4, further comprising: In the case where the point cloud data in the channel is clustered to obtain a non-vehicle object, monitoring the duration of the existence of the non-vehicle object; When the duration is greater than the time threshold, an early warning is issued.

6. The vehicle detection method according to claim 1, wherein: Determining at least one of the position information and the speed information of each vehicle to be inspected according to the vehicle point cloud data includes: Determine at least one of the front position, rear position and speed of the vehicle to be inspected; or Determining the position information of the load of each vehicle to be inspected according to the target point cloud data includes: Determine the location of the front and rear cargo compartments of each vehicle to be inspected.

7. The vehicle detection method according to any one of claims 1 to 6, wherein: The load includes a load in a box structure.

8. A vehicle detection device, comprising: A first acquisition module is configured to acquire original point cloud data using a multi-line laser radar when at least one vehicle to be inspected is traveling on a detection section, and determine a laser scanning line to which each point cloud of each vehicle to be inspected in the original point cloud data belongs, wherein the multi-line laser radar is installed on the side of the detection section for scanning the at least one vehicle to be inspected; A second acquisition module is configured to acquire vehicle point cloud data of each of the at least one vehicle to be inspected according to the original point cloud data; A first determination module is configured to determine at least one of the position information and the speed information of each vehicle to be inspected according to the vehicle point cloud data; A second determination module is configured to determine, from the vehicle point cloud data, target point cloud data of each vehicle to be inspected obtained by scanning with at least one laser line having a vertical field of view angle greater than a vertical field of view angle threshold; The third determination module is configured to determine the position information of the load of each vehicle to be inspected according to the target point cloud data of each vehicle to be inspected.

9. A vehicle detection device, comprising: Memory; as well as A processor coupled to the memory, wherein the processor is configured to execute the vehicle detection method according to any one of claims 1 to 7 based on instructions stored in the memory.

10. A vehicle detection system comprising: The vehicle detection device as claimed in claim 8 or 9; The multi-line laser radar is configured to scan at least one vehicle to be inspected traveling on a detection section, obtain original point cloud data of the at least one vehicle to be inspected, and send the original point cloud data to the vehicle detection device.

11. A computer-readable storage medium having computer program instructions stored thereon, wherein the instructions, when executed by a processor, implement the vehicle detection method according to any one of claims 1 to 7.

Images