Radar step recognition method and system

Abstract

The invention provides a radar step recognition method and system, and the method comprises the steps: firstly obtaining a preset calibration point coordinate which is point cloud data recorded according to a standard step; acquiring and acquiring coordinates of scanning points by using a scanning unit; calculating a registration matrix through preset calibration point coordinates and collected scanning point coordinates; a registration coordinate system is generated according to the collected scanning point coordinates and the registration matrix, and the registration coordinate system and a coordinate system of the preset calibration point coordinates are the same coordinate system; finally, the number of falling point clouds of the acquisition scanning points in a preset range in the registration coordinate system is calculated; and if the number of the falling point clouds is greater than or equal to the point cloud number threshold value, outputting a first identification result for representing that the pace is trampled. According to the method, the coordinate systems of the preset calibration points and the acquisition scanning points are registered, and then the number of point clouds of the acquisition scanning points falling into the preset range is judged, so that whether the pace is treaded or not is identified, and the problem of inaccurate position point identification is solved.

Description

Technical Field

[0001] The present application relates to the field of target recognition, and in particular to a radar step recognition method and system. Background Art

[0002] In order to increase the safety of train operation, locomotives and vehicles need to be inspected regularly. In order to increase the standardization of train inspections, the relevant railway departments have formulated the Inspection Law, which stipulates the inspection steps, inspection sequence, key inspection parts, components and quality standards and requirements.

[0003] The inspection method includes the 25-step inspection method, which is to have inspectors check different parts and components of the train at different locations, as well as the quality standard requirements of the components. Different devices can be used to detect the location points during the inspection.

[0004] Fixed foot switches and mobile foot switches are set at different locations, and the inspector is judged by stepping on the location point. The fixed foot switch needs to be buried underground and cannot be moved. Although the mobile foot switch can be moved, it cannot be fixed on the location point and is easy to move, resulting in inaccurate location point identification. Summary of the invention

[0005] The present application provides a radar step recognition method and system to solve the problem of inaccurate position point recognition.

[0006] In a first aspect, the present application provides a radar step recognition method, comprising:

[0007] Obtaining the coordinates of the preset calibration points, wherein the preset calibration point coordinates are point cloud data recorded according to the standard steps;

[0008] Using the scanning unit to obtain the coordinates of the collected scanning points;

[0009] Calculating a registration matrix by using the preset calibration point coordinates and the collected scanning point coordinates;

[0010] Generate a registration coordinate system according to the acquired scanning point coordinates and the registration matrix, wherein the registration coordinate system is the same coordinate system as the coordinate system of the preset calibration point coordinates;

[0011] Calculate the number of point clouds that fall within a preset range of the acquisition scanning points in the registration coordinate system;

[0012] If the number of the fallen point clouds is greater than or equal to the point cloud number threshold, a first recognition result is output, and the first recognition result is used to characterize that the step is hit.

[0013] In some embodiments of the present application, obtaining the coordinates of the preset calibration points includes:

[0014] Acquire a target path and preset calibration points, wherein the preset calibration points are set on the target path;

[0015] Calculating a first point cloud density and a second point cloud density of the preset calibration point, wherein the first point cloud density is the point cloud density of the preset calibration point at a first time, and the second point cloud density is the point cloud density of the preset calibration point at a second time;

[0016] If the second point cloud density is greater than the first point cloud density, the coordinates of the preset calibration point are obtained.

[0017] In some embodiments of the present application, the acquiring the coordinates of the calibration preset point and the acquiring the coordinates of the scanning point includes:

[0018] Setting target point information, wherein the target point information includes the number of the target point and the detection information of the target point;

[0019] According to the target point information, the preset calibration point coordinates and the acquisition scanning point coordinates are acquired, and the target point information has a mapping relationship with the preset calibration point coordinates and the acquisition scanning point coordinates.

[0020] In some embodiments of the present application, the method further includes:

[0021] Acquire first target point information and second target point information;

[0022] According to the first target point information and the second target point information, the first preset calibration point coordinates, the second preset calibration point coordinates, the first acquisition scanning point coordinates and the second acquisition scanning point coordinates are obtained, the first preset calibration point coordinates are preset calibration point coordinates having a mapping relationship with the first target point information, the second preset calibration point coordinates are preset calibration point coordinates having a mapping relationship with the second target point information, the first acquisition scanning point coordinates are acquisition scanning point coordinates having a mapping relationship with the first target point information, and the second acquisition scanning point coordinates are acquisition scanning point coordinates having a mapping relationship with the second target point information.

[0023] In some embodiments of the present application, the calculating the registration matrix by using the coordinates of the preset calibration points and the coordinates of the collected scanning points further includes:

[0024] Obtaining a rotation angle, where the rotation angle is the angle between a line connecting the first preset calibration point coordinates and the first acquisition scanning point coordinates, and a line connecting the second preset point calibration coordinates and the second acquisition scanning point coordinates;

[0025] Calculate the rotation matrix through the rotation angle;

[0026] Obtain a first translation matrix and a second translation matrix, wherein the first translation matrix is ​​a matrix obtained by translating the coordinates of the first target point to an origin, and the second translation matrix is ​​a matrix obtained by translating the origin to the coordinates of the second target point;

[0027] A registration matrix is ​​calculated, where the registration matrix is ​​the product of the rotation matrix, the first translation matrix, and the second translation matrix.

[0028] In some embodiments of the present application, generating a registration coordinate system according to the acquired scanning point coordinates and the registration matrix includes:

[0029] Sending a scanning instruction to the scanning unit to control the scanning unit to scan and collect scanning point coordinates in real time;

[0030] Calculating transformed coordinates, wherein the transformed coordinates are the product of the acquisition scanning point coordinates and the registration matrix;

[0031] A registration coordinate system is generated, wherein the registration coordinate system is a coordinate system of the converted coordinates and a coordinate system of the calibration point coordinates.

[0032] In some embodiments of the present application, calculating the number of point clouds that fall within a preset range of the acquisition scanning point in the registration coordinate system includes:

[0033] Acquire a target image of continuous frames, wherein the target image is an image with acquisition scanning points;

[0034] Setting a preset range, wherein the preset range is a preset radius range centered at the preset calibration point in the registration coordinate system;

[0035] The number of point clouds of the acquired scanning points in the target image falling within the preset range is calculated.

[0036] In some embodiments of the present application, if the number of falling point clouds is greater than or equal to the point cloud number threshold, outputting a first recognition result includes:

[0037] Setting a point cloud quantity threshold according to the angular resolution of the scanning unit;

[0038] Setting a frame number threshold according to a scanning frequency of the scanning unit;

[0039] If the number of point clouds falling into the target image equal to the frame number threshold is greater than or equal to the point cloud number threshold, a first recognition result is output.

[0040] In some embodiments of the present application, the method further includes:

[0041] If the number of the falling point clouds is less than the point cloud number threshold, a second recognition result is output, where the second recognition result is used to represent one or more of a missed step, a missed step, and a misplaced step;

[0042] A scanning instruction is sent to the scanning unit according to the second recognition result to control the scanning unit to scan and collect scanning points again.

[0043] In a second aspect, the present application provides a radar step recognition system, which is applied to the radar step recognition method according to any one of the first aspects of the claims, and the system includes: an acquisition module, a scanning module, a registration module and a calculation module;

[0044] The acquisition module is used to acquire preset calibration points, wherein the coordinates of the preset calibration points are point cloud data recorded according to standard steps;

[0045] The scanning module is used to obtain acquisition scanning points;

[0046] The scanning module is also used to obtain the coordinates of the preset point and the coordinates of the acquisition scanning point;

[0047] The registration module is used to generate a registration matrix through the coordinates of the preset calibration points and the coordinates of the acquisition scanning points; and is also used to generate a registration coordinate system according to the acquisition scanning point coordinates and the registration matrix, wherein the registration coordinate system is the same coordinate system as the coordinate system of the preset calibration point coordinates;

[0048] The calculation module is used to calculate the number of point clouds that the collected scanning points fall into within a preset range in the registration coordinate system. If the number of point clouds that fall into the point clouds is greater than or equal to a point cloud number threshold, a first recognition result is output, and the first recognition result is used to characterize that a step has been hit.

[0049] It can be seen from the above technical solutions that the present application provides a radar step recognition method and system, the method first obtains the preset calibration point coordinates, wherein the preset calibration point coordinates are point cloud data recorded according to the standard steps; then uses the scanning unit to obtain the acquisition scanning point coordinates; calculates the registration matrix through the preset calibration point coordinates and the acquisition scanning point coordinates; generates a registration coordinate system according to the acquisition scanning point coordinates and the registration matrix, wherein the registration coordinate system is the same coordinate system as the coordinate system of the preset calibration point coordinates; finally calculates the number of point clouds that fall within the preset range of the acquisition scanning point in the registration coordinate system; if the number of falling point clouds is greater than or equal to the point cloud number threshold, outputs a first recognition result for characterizing that the step is stepped on. By aligning the coordinate systems of the preset calibration points and the acquisition scanning points, and then determining the number of point clouds that fall within the preset range of the acquisition scanning points, it is possible to identify whether the step is stepped on, so as to solve the problem of inaccurate position point recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solution of the present application, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0051] Figure 1 The following is a schematic diagram showing the flow of the radar identification method for this embodiment;

[0052] Figure 2 This is a schematic diagram of the path shown in this embodiment;

[0053] Figure 3 This is a schematic diagram of outputting results according to the number of point clouds shown in this embodiment;

[0054] Figure 4 The structure diagram of the radar recognition system is shown in the present embodiment.

[0055] Illustration Description:

[0056] Among them, 100 is an acquisition module, 200 is a scanning module, 300 is a registration module, and 400 is a calculation module. DETAILED DESCRIPTION

[0057] The following embodiments are described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following embodiments do not represent all implementations consistent with the present application. They are only examples of systems and methods consistent with some aspects of the present application as detailed in the claims.

[0058] In some embodiments, during train inspection, the position point can be identified by a fixed foot switch and a mobile foot switch, wherein the fixed foot switch is a foot pedal, which includes an upper and lower part, the upper part is an upper cover, and the lower part is buried at the position point, and the upper part is connected to the lower part by a spring. The fixed foot switch determines whether it is stepped on by transmitting a signal through a travel switch. When the upper cover is not stepped on, the upper cover is lifted up by the spring, and the travel switch in the lower part is in a non-contact state and does not transmit a signal. When the upper cover is stepped on, the spring is compressed, the travel switch is in a contact state, and a signal is transmitted to detect the step position. However, the fixed foot switch requires a dedicated site, and a groove is opened on the site. Moreover, due to the limited site area, it cannot support a variety of vehicle inspection methods. The foot switch is buried underground, which is not convenient for maintenance.

[0059] The mobile foot switch includes a cover plate, a conductive upper cover plate, an isolating switch, a conductive lower cover plate and a bottom plate. When stepped on, the conductive upper cover plate contacts the conductive lower cover plate. When not stepped on, the conductive upper cover plate does not contact the conductive lower cover plate. Whether the position point is stepped on is determined by whether the conductive upper cover plate contacts the conductive lower cover plate. After contact, the contact signal is transmitted through the wireless communication module. Since the inspection method has many steps and requires a large number of switches, the maintenance workload is large, for example, the charging workload is large, and since it is not buried underground, the switch is difficult to fix. During the inspection process, it is easy to be moved, resulting in inaccurate position point identification.

[0060] To solve the above problem of inaccurate position point recognition, see Figure 1 Some embodiments of the present application provide a radar step recognition method, including:

[0061] S100: Obtaining the coordinates of preset calibration points.

[0062] In this embodiment, the preset calibration point coordinates are point cloud data recorded according to the standard pace, the standard pace represents the pace of walking according to the pace prescribed by the vehicle inspection method during the vehicle inspection method training process, and each point cloud data represents a preset calibration point coordinate.

[0063] For example, see Figure 2 During the vehicle inspection training, if the vehicle inspection method is the 25-step vehicle inspection method, Figure 2 1-25 represents 25 steps, and the coordinates of the preset calibration points are the coordinates of 1-25 standard steps. The coordinates of each preset calibration point are different. In the 25-step vehicle inspection method, the number of coordinates of the preset calibration points is 25.

[0064] In some embodiments, to obtain the coordinates of the preset calibration points, the target path and the preset calibration points can be first obtained; then the first point cloud density and the second point cloud density of the preset calibration points are calculated; if the second point cloud density is greater than the first point cloud density, the coordinates of the preset calibration points are obtained.

[0065] The target path may be one or more to detect different positions of the vehicle. For example, the vehicle inspection method includes two target paths. The first target path is used to detect the front rails and the front of the vehicle, and the second target path is used to detect the rear rails and the rear of the vehicle.

[0066] There are multiple different preset calibration points on different target paths, each preset calibration point represents a different step point coordinate, and different preset calibration points can detect different parts and components of the vehicle.

[0067] In order to ensure that the coordinates of the preset calibration points are consistent with the coordinates specified in the vehicle inspection method, you can first walk according to the pace requirements specified in the vehicle inspection method. When you walk to each preset calibration point, pause at the preset calibration point. The pause time can be determined based on the growth rate of the point cloud density of the preset calibration point. In this embodiment, pause at the preset calibration point for three seconds to increase the point cloud density of the preset calibration point.

[0068] The preset calibration point can be obtained by the point cloud density of the preset calibration point at different times. For example, before walking, it is set to the first time, and the point cloud density at the first time is the first point cloud density. Since there is no walking path, the first point cloud density is 0. During walking, it is set to the second time. It can be understood that the second time may be multiple times, for example: when walking to the first preset calibration point, the time of obtaining the point cloud density; for another example: after walking, the time of obtaining the point cloud density; the point cloud density obtained at the second time is the second point cloud density. If there is a second point cloud density greater than the first point cloud density at the preset calibration point, this step point is set as the preset calibration point.

[0069] In the process of walking from the first step to the second step, before walking, the point cloud density of the first step and the second step is obtained, and the point cloud density of the first step and the second step can be obtained after the first step. After passing the first step, the point cloud density of the first step increases, and the size of the point cloud density can be determined according to the dwelling time; the point cloud density of the first step and the second step can also be obtained after passing the first step and the second step, and the point cloud density of the first step and the second step both increase.

[0070] S200: using a scanning unit to acquire the coordinates of the scanning points.

[0071] When the scanning unit is used in different places or used in the same place but scans again after charging, the position and scanning angle will change, and alignment is required to make the coordinate system of the scanning unit scanning coordinates consistent with the coordinate system of the preset calibration point.

[0072] The scanning unit in this embodiment is a single-line radar. In some embodiments, it also includes a wifi communication unit. The wifi communication unit and the single-line radar are connected by a network cable. The single-line radar includes a laser transmitter and a 360-degree rotating scanner. By transmitting a detection signal (laser beam) to the target and comparing the received signal reflected from the target (target echo) with the transmitted signal, the scanning result of the target can be obtained, such as target distance, direction, height, speed, posture, and even shape parameters.

[0073] For example, when a single-line radar needs to obtain the coordinates of the acquisition points, that is, to obtain point cloud data, the single-line radar obtains the scanning result data, and can filter, parse, and process the scanning result data to generate point cloud data. In some embodiments, these point cloud data can also be visualized, for example, using a visualization tool PLC (Point Cloud Library). Through visualization, these point cloud data can be converted into graphics, images, and other forms for easy observation.

[0074] In order to align the coordinate systems of the collected scan points and the preset calibration points, the ICP (Iterative Closest Point) algorithm can be used for alignment. The ICP algorithm can align the coordinate systems of two or more point clouds. The ICP algorithm includes the coordinate system alignment of two or more point clouds, for example; four-point alignment. The four-point alignment finds four coplanar corresponding points and calculates the intersection ratio so that it remains unchanged under affine transformation. It can be seen that the four-point alignment has high requirements for the selection of points. Therefore, in this embodiment, the two-point alignment method is used for alignment, that is, two points at the same position are selected. For these two points, the coordinates of the preset calibration point and the coordinates of the collected scan points are obtained, and then the coordinate system is aligned.

[0075] In order to obtain the coordinates of the preset calibration point and the coordinates of the acquisition scanning point as the same step point, in some embodiments, the target point information can be set, and then the preset calibration point coordinates and the acquisition scanning point coordinates can be obtained based on the target point information.

[0076] The target point information includes the number of the target point and the detection information of the target point. The number of the target point can be B1, B2, B3, ..., Bn. The detection information of the target point is the information of the train parts and the position of the train checked at point B1. For example, the target point information includes the number of the target point B2, and the detection information of the target point is the detection of the coupler joist, the coupler tail frame, the buffer, etc. It can be understood that the registration method used in this embodiment is the two-point registration method, and the target point information can be the numbers of the two target points, for example, the target point numbers are selected as B2 and B3.

[0077] The target point information has a mapping relationship with the preset calibration point coordinates and the acquisition scanning point coordinates, that is, the unique target point information corresponds to the unique preset calibration point coordinates, and the unique target point information corresponds to the unique acquisition scanning point coordinates. Exemplarily, the target point information includes the target point number B2, the preset calibration point coordinates a1(x n ,y n ), the acquisition scanning point coordinates are a2(x n1 ,y n1 ), B2 corresponds to unique a1 and a2. It can be understood that the preset calibration point coordinates are not associated with x and y in the acquisition scanning point coordinates, and are only examples.

[0078] S300: Calculate the registration matrix by presetting the coordinates of the calibration points and collecting the coordinates of the scanning points.

[0079] According to the above, this embodiment uses the two-point registration method for registration. In order to obtain the preset calibration points and acquisition scanning points of the same position point, in some embodiments, the first target point information and the second target point information can be obtained first, and then the first preset calibration point coordinates, the second preset calibration point coordinates, the first acquisition scanning point coordinates and the second acquisition scanning point coordinates can be obtained based on the first target point information and the second target point information.

[0080] Among them, the first preset calibration point coordinates are preset calibration point coordinates having a mapping relationship with the first target point information, the second preset calibration point coordinates are preset calibration point coordinates having a mapping relationship with the second target point information, the first acquisition scanning point coordinates are acquisition scanning point coordinates having a mapping relationship with the first target point information, and the second acquisition scanning point coordinates are acquisition scanning point coordinates having a mapping relationship with the second target point information.

[0081] It is understandable that the first target point information and the second target point information are obtained randomly in the standard step. For example, if the number of the first target point is selected as B2 and the number of the second target point is selected as B3, according to the mapping relationship, the coordinates of the first preset calibration point are o1(x 2 ,y 2 ), the coordinates of the second preset calibration point are o2(x 3 ,y 3 ), the coordinates of the first acquisition scanning point are n1(x 21 ,y 21 ), the coordinates of the second acquisition scanning point are n2(x 31 ,y 31 ).

[0082] After obtaining the coordinates of two preset calibration points and the coordinates of two collected scanning points, a registration matrix can be generated to align the coordinate system. In some embodiments, the rotation angle is obtained and the rotation matrix is ​​calculated by the rotation angle; the first translation matrix and the second translation matrix are obtained.

[0083] The rotation angle is the angle between the line connecting the first preset calibration point coordinates and the first acquisition scanning point coordinates, and the angle between the line connecting the second preset point calibration coordinates and the second acquisition scanning point coordinates, that is, the angle between the line connecting o1 and n1 and the line connecting o2 and n2. Exemplarily, o1 can be rotated counterclockwise around n1 so that the two lines are parallel, and the direction of the line connecting n1 and o1 is the same as the direction of the line connecting n2 and o2. The calculated rotation angle is the rotation angle.

[0084] After obtaining the rotation angle, calculate the rotation matrix of the rotation angle and obtain the rotation matrix MR 3*3 .

[0085] The first translation matrix is ​​the matrix obtained by translating the coordinates of the first target point to the origin, and the second translation matrix is ​​the matrix obtained by translating the origin to the coordinates of the second target point; translate n1 to the origin (0,0) to obtain the translation matrix MT1 3*3 , which is the first translation matrix; translate the origin (0,0) to n2 to obtain the translation matrix MT2 3*3 , which is the second translation matrix.

[0086] After obtaining the rotation matrix, the first translation matrix and the second translation matrix, the registration matrix can be calculated. 3*3 , translation matrix MT1 3*3 And the translation matrix MT2 3*3 The product of 3*3 .

[0087] It is understandable that when the scanning unit is used at a different location or is put back after being charged, it needs to be re-registered to obtain a new registration matrix to align the coordinate system.

[0088] S400: Generate a registration coordinate system according to the acquired scanning point coordinates and the registration matrix.

[0089] After the registration matrix is ​​generated, the coordinate system of the acquired scanning points can be registered, wherein the registration coordinate system is the same coordinate system as the coordinate system of the preset calibration point coordinates.

[0090] To generate a registration coordinate system, in some embodiments, a scanning instruction is first sent to the scanning unit to control the scanning unit to scan and collect scanning point coordinates in real time; then the converted coordinates are calculated, wherein the converted coordinates are the product of the collected scanning point coordinates and the registration matrix; and a registration coordinate system is generated.

[0091] It can be understood that the registration coordinate system is the coordinate system of the converted coordinates and the coordinate system of the preset calibration point coordinates, that is, the registration coordinate system is the same as the acquisition scanning point coordinate system and the preset calibration point coordinate system, and the acquisition scanning point coordinate system is registered to the preset calibration point coordinate system.

[0092] S500: Calculate the number of point clouds that fall within a preset range in the registration coordinate system.

[0093] After the coordinate system is aligned, the collected scanning points are reflected in the aligned coordinate system. The falling point cloud data of the collected scanning points can be calculated on the aligned coordinate system to calculate whether the step is hit.

[0094] To calculate the number of point clouds, a preset range is first set. In some embodiments, by acquiring continuous frames of target images and then setting the preset range, the number of point clouds whose collected scanning points in the target image fall within the preset range is calculated.

[0095] The target image can be obtained after processing the information received by the scanning unit. The target image is an image with acquisition scanning points. The target image can be an image with one acquisition scanning point or an image with multiple acquisition scanning points. Therefore, the number of target images can be one or more. However, it should be noted that when the number of target images is multiple, the number of each acquisition scanning point in the multiple target images is one, that is, all acquisition scanning points in the multiple target images are calculated only once, thereby reducing repeated calculations.

[0096] It is understandable that after acquiring the target image, each frame of the image may be preprocessed for easy recognition, including operations such as denoising and binarization.

[0097] The preset range is a preset radius range centered at the preset calibration point in the registration coordinate system. In this embodiment, the preset radius is 25 cm.

[0098] Exemplarily, the target image includes multiple ones, the first target image has one acquisition scanning point, the target point mapped by one acquisition scanning point is numbered B1, and the preset calibration point mapped by B1 is taken as the center of the circle. The number of point clouds within a radius of 25 cm is calculated, and the number of point clouds obtained is n.

[0099] S510: If the number of point clouds falling into the point cloud is greater than or equal to the point cloud number threshold, output a first recognition result.

[0100] The number of point clouds is obtained according to the above calculation. If the number of point clouds is greater than or equal to the point cloud number threshold, it means that the step is hit, wherein the first recognition result is used to characterize that the step is hit.

[0101] In the process of collecting target images, some stepping behaviors may be missed due to the sampling frequency or the instantaneous nature of the action. For example, if a preset calibration point is only stepped on briefly but this stepping behavior is not captured in a single frame image, then this stepping behavior will be ignored, affecting the accuracy of the calculation. Therefore, in addition to setting the point cloud number threshold, in this embodiment, a frame number threshold is also set.

[0102] In some embodiments, a point cloud quantity threshold is first set according to the angular resolution of the scanning unit; then a frame number threshold is set according to the scanning frequency of the scanning unit; if the number of point clouds falling into the target image equal to the frame number threshold is greater than or equal to the point cloud quantity threshold, a first recognition result is output.

[0103] The frame number threshold and the point cloud number threshold are determined according to the scanning frequency and angular resolution of the scanning unit. If the scanning frequency of the scanning unit is 25 Hz, the frame number threshold is 4. If the angular resolution of the scanning unit is 0.25°, the point cloud number threshold is 3.

[0104] Exemplarily, when calculating the number of point clouds that fall into a target image, first determine how many acquisition scanning points exist in the target image. If the number of acquisition scanning points is 2, then obtain the numbers of the target points mapped by the two acquisition scanning points, and determine the preset range by number. Within a radius of 25 cm between two preset calibration points, calculate the number of point clouds within two preset radius ranges in four consecutive image frames. If the number of point clouds is greater than or equal to 3, then it is considered that the step has been hit.

[0105] In the process of calculating the number of point clouds, in addition to the steps being stepped on, there are other situations. To detect other situations, see Figure 3 If the number of points falling into the point cloud is less than the point cloud number threshold, a second recognition result is output, wherein the second recognition result is used to characterize one or more of a step not being stepped on, a step being missed, and a step being misplaced; and then a scanning instruction is sent to the scanning unit according to the second recognition result to control the scanning unit to scan and collect scanning points again.

[0106] By judging that the number of point clouds is less than the point cloud number threshold, another recognition result is output, indicating that the step is not stepped on, the step is missed, the step is misplaced, etc. After the above situation occurs, the scanning unit can be controlled to scan again to confirm whether the recognition result is accurate, so as to improve the recognition accuracy.

[0107] Exemplarily, within the preset radius, the number of point clouds is 2, and a second recognition result is output. It can be understood that when the second recognition result appears, there is no need to wait for the completion of the judgment of the number of all point clouds, and a command for another scan is directly sent to the scanning unit.

[0108] Based on the above-mentioned radar step recognition method, the present application embodiment also provides a radar step recognition system, see Figure 4 The system includes: an acquisition module 100, a scanning module 200, a registration module 300 and a calculation module 400.

[0109] The acquisition module 100 is used to acquire preset calibration points, where the coordinates of the preset calibration points are point cloud data recorded according to the standard steps;

[0110] The scanning module 200 is used to obtain the collection scanning points;

[0111] The scanning module 200 is also used to obtain the coordinates of the preset points and collect the coordinates of the scanning points;

[0112] The registration module 300 is used to generate a registration matrix by using the coordinates of the preset calibration points and the coordinates of the collected scanning points; and is also used to generate a registration coordinate system according to the coordinates of the collected scanning points and the registration matrix, wherein the registration coordinate system is the same coordinate system as the coordinate system of the preset calibration point coordinates;

[0113] The calculation module 400 is used to calculate the number of point clouds that fall within a preset range in the registration coordinate system. If the number of point clouds that fall within is greater than or equal to a point cloud number threshold, a first recognition result is output. The first recognition result is used to characterize that a step has been hit.

[0114] In some embodiments, the system further includes a power supply, which is connected to the scanning unit and is used to provide power to the scanning module 200. In this embodiment, the power supply may be a rechargeable battery. It is understood that when the scanning module 200 is used in different venues or used in the same venue but scanned again after charging, the position and scanning angle will change, and alignment is required to make the coordinate system of the scanning unit scanning coordinates consistent with the coordinate system of the preset calibration point. Therefore, after the scanning module 200 is charged with the rechargeable battery, the coordinate system needs to be re-aligned. It is understood that the scanning module 200 and the above-mentioned scanning unit are both single-line radars.

[0115] It can be seen from the above technical solutions that the embodiment of the present application provides a radar step recognition method and system, the method first obtains the preset calibration point coordinates, wherein the preset calibration point coordinates are point cloud data recorded according to the standard steps; then uses the scanning unit to obtain the acquisition scanning point coordinates; calculates the registration matrix through the preset calibration point coordinates and the acquisition scanning point coordinates; generates a registration coordinate system according to the acquisition scanning point coordinates and the registration matrix, wherein the registration coordinate system is the same coordinate system as the coordinate system of the preset calibration point coordinates; finally calculates the number of point clouds that fall within the preset range of the acquisition scanning point in the registration coordinate system; if the number of point clouds that fall within is greater than or equal to the point cloud number threshold, outputs a first recognition result for characterizing that the step is stepped on. By aligning the coordinate systems of the preset calibration points and the acquisition scanning points, and then determining the number of point clouds that fall within the preset range of the acquisition scanning points, it is possible to identify whether the step is stepped on, so as to solve the problem of inaccurate position point recognition.

[0116] Similar parts between the embodiments provided in this application can be referenced to each other. The specific implementation methods provided above are only a few examples under the general concept of this application and do not constitute a limitation on the protection scope of this application. For those skilled in the art, any other implementation methods expanded based on the scheme of this application without creative work belong to the protection scope of this application.

Claims

1. A radar recognition step method, It is characterized in that include: Obtaining the coordinates of the preset calibration points, wherein the preset calibration point coordinates are point cloud data recorded according to the standard steps; Using the scanning unit to obtain the coordinates of the collected scanning points; Calculate the registration matrix by using the preset calibration point coordinates and the collected scanning point coordinates; Generate a registration coordinate system according to the acquired scanning point coordinates and the registration matrix, wherein the registration coordinate system is the same coordinate system as the coordinate system of the preset calibration point coordinates; Calculate the number of point clouds that fall within a preset range of the acquisition scanning points in the registration coordinate system; If the number of the fallen point clouds is greater than or equal to the point cloud number threshold, a first recognition result is output, and the first recognition result is used to characterize that the step is hit.

2. The radar identification method according to claim 1, It is characterized in that The step of obtaining the coordinates of the preset calibration points includes: Acquire a target path and preset calibration points, wherein the preset calibration points are set on the target path; Calculating a first point cloud density and a second point cloud density of the preset calibration point, wherein the first point cloud density is the point cloud density of the preset calibration point at a first time, and the second point cloud density is the point cloud density of the preset calibration point at a second time; If the second point cloud density is greater than the first point cloud density, the coordinates of the preset calibration point are obtained.

3. The radar identification method according to claim 1, It is characterized in that The method further comprises: Setting target point information, wherein the target point information includes the number of the target point and the detection information of the target point; According to the target point information, the preset calibration point coordinates and the acquisition scanning point coordinates are acquired, and the target point information has a mapping relationship with the preset calibration point coordinates and the acquisition scanning point coordinates.

4. The radar identification method according to claim 3, It is characterized in that The calculating the registration matrix by using the coordinates of the preset calibration points and the coordinates of the collected scanning points includes: Acquire first target point information and second target point information; According to the first target point information and the second target point information, the first preset calibration point coordinates, the second preset calibration point coordinates, the first acquisition scanning point coordinates and the second acquisition scanning point coordinates are obtained, the first preset calibration point coordinates are preset calibration point coordinates having a mapping relationship with the first target point information, the second preset calibration point coordinates are preset calibration point coordinates having a mapping relationship with the second target point information, the first acquisition scanning point coordinates are acquisition scanning point coordinates having a mapping relationship with the first target point information, and the second acquisition scanning point coordinates are acquisition scanning point coordinates having a mapping relationship with the second target point information.

5. The radar identification method according to claim 4, It is characterized in that The calculation of the registration matrix by using the coordinates of the preset calibration points and the coordinates of the collected scanning points also includes: Obtaining a rotation angle, where the rotation angle is the angle between a line connecting the first preset calibration point coordinates and the first acquisition scanning point coordinates, and a line connecting the second preset point calibration coordinates and the second acquisition scanning point coordinates; Calculate the rotation matrix through the rotation angle; Obtain a first translation matrix and a second translation matrix, wherein the first translation matrix is ​​a matrix obtained by translating the coordinates of the first target point to an origin, and the second translation matrix is ​​a matrix obtained by translating the origin to the coordinates of the second target point; A registration matrix is ​​calculated, where the registration matrix is ​​the product of the rotation matrix, the first translation matrix, and the second translation matrix.

6. The radar identification method according to claim 1, It is characterized in that The generating a registration coordinate system according to the acquired scanning point coordinates and the registration matrix comprises: Sending a scanning instruction to the scanning unit to control the scanning unit to scan and collect scanning point coordinates in real time; Calculating transformed coordinates, wherein the transformed coordinates are the product of the acquisition scanning point coordinates and the registration matrix; A registration coordinate system is generated, wherein the registration coordinate system is a coordinate system of the converted coordinates and a coordinate system of the calibration point coordinates.

7. The radar identification method according to claim 1, It is characterized in that The calculating the number of point clouds that the acquired scanning points fall into within a preset range in the registration coordinate system includes: Acquire a target image of continuous frames, wherein the target image is an image with acquisition scanning points; Setting a preset range, wherein the preset range is a preset radius range centered at the preset calibration point in the registration coordinate system; The number of point clouds of the acquired scanning points in the target image falling within the preset range is calculated.

8. The radar identification method according to claim 7, It is characterized in that If the number of the falling point clouds is greater than or equal to the point cloud number threshold, outputting a first recognition result includes: Setting a point cloud quantity threshold according to the angular resolution of the scanning unit; Setting a frame number threshold according to a scanning frequency of the scanning unit; If the number of point clouds falling into the target image equal to the frame number threshold is greater than or equal to the point cloud number threshold, a first recognition result is output.

9. The radar identification method according to claim 1, It is characterized in that The method further comprises: If the number of the falling point clouds is less than the point cloud number threshold, a second recognition result is output, where the second recognition result is used to represent one or more of a missed step, a missed step, and a misplaced step; A scanning instruction is sent to the scanning unit according to the second recognition result to control the scanning unit to scan and collect scanning points again.

10. A radar step recognition system, It is characterized in that The radar step recognition method according to any one of claims 1 to 9, wherein the system comprises: An acquisition module, the acquisition module is used to acquire preset calibration points, the coordinates of the preset calibration points are point cloud data recorded according to standard steps; A scanning module, the scanning module is used to obtain acquisition scanning points; The scanning module is also used to obtain the coordinates of the preset point and the coordinates of the acquisition scanning point; A registration module, the registration module is used to generate a registration matrix through the coordinates of the preset calibration points and the coordinates of the acquisition scanning points; and is also used to generate a registration coordinate system according to the acquisition scanning point coordinates and the registration matrix, the registration coordinate system and the coordinate system of the preset calibration point coordinates are the same coordinate system; A calculation module is used to calculate the number of point clouds that the collected scanning points fall into within a preset range in the registration coordinate system. If the number of point clouds that fall into the point clouds is greater than or equal to a point cloud number threshold, a first recognition result is output. The first recognition result is used to characterize that a step has been hit.

Images