Vehicle inspection step recognition method and system based on single-line laser radar

Abstract

The present invention discloses a vehicle inspection step recognition method and system based on a single-line laser radar, which relates to the field of laser radar sensor recognition technology, including: obtaining point cloud data for each scan; generating a circular area with a preset radius with each standard step position as the center of the circle; judging whether the number of data points of the point cloud data scanned more than twice in a row that fall into a certain circular area exceeds a predetermined number, if so, the standard step position corresponding to the circular area is identified as being stepped on, otherwise it is identified as not being stepped on; comparing the state change of each standard step position with the step position and step sequence specified in the vehicle inspection method, and outputting the recognition result of the vehicle inspection step. The method of the present invention can accurately identify the position and sequence of the steps.

Description

Technical Field

[0001] The present invention relates to the field of laser radar sensor recognition technology, and in particular to a vehicle inspection step recognition method and system based on a single-line laser radar. Background Art

[0002] In order to ensure the safety of railway locomotives, locomotives need to be inspected regularly. To ensure the standardization of inspections, the relevant railway departments have formulated inspection laws, which stipulate detailed inspection steps, inspection sequence, key parts, components and quality standards. At the same time, the inspection law also stipulates the order and position of the inspectors' steps. Figure 1 This is a schematic diagram of the 25-step inspection method for the C64K(H) high-speed railway freight car. Strictly following the inspection steps, step positions and step sequence in the diagram can ensure the standardization of locomotive inspection, thereby avoiding missing inspection points and minimizing safety hazards. Therefore, in order to make inspectors strictly follow the above inspection method to perform locomotive maintenance, it is necessary to conduct strict and standardized training for inspectors. During the training process, step detection and judging the training effect or correcting the wrong actions during the training process based on the detected step position sequence have become technical issues that must be solved by the locomotive maintenance training system.

[0003] There are two main methods of step detection technology in the prior art:

[0004] Method 1: Fixed mechanical foot switch. Figure 2 , dig a pit at the step position, embed the foot switch, then make a groove on the ground, lay the foot switch signal line in it, and transmit the signal of whether the foot switch is stepped on to the background signal processing system through the signal line. When the upper cover of the foot switch is not stepped on, it is lifted by the spring, and the internal travel switch is in a non-contact state; when the foot switch is stepped on, the internal spring is compressed, and the internal travel switch is in a contact state; the contact state of the internal travel switch is uploaded to the background signal processing system in real time, thereby realizing the detection of the current step position. However, the disadvantages of this technical solution are: it requires a fixed dedicated practice site, the amount of engineering work is large, pits need to be dug and grooves need to be made on the ground for wiring, it does not support multiple vehicle inspection methods, and supports at most two vehicle inspection methods, that is, a set of vehicle inspection method foot switches are installed on both sides, and in addition, the foot switch is buried underground, and the daily maintenance workload is large.

[0005] Method 2: Movable foot switch. Figure 3The principle of this scheme is: when the foot is on the foot switch, the conductive upper cover plate and the conductive lower cover plate are in contact; when the foot is not on the foot switch, the conductive upper cover plate and the conductive lower cover plate are not in contact. The foot switch is judged by judging whether the upper and lower covers are in contact to determine whether it is stepped on. The contact signal of the upper and lower covers is uploaded to the background signal processing system through the wireless communication module for processing. However, the disadvantages of this technical scheme are: the switch is difficult to fix on the ground and is easy to be kicked away. The inspector is prone to slipping and falling when stepping on it, which affects the training of the inspector; frequent use will cause the conductive upper cover plate, the conductive lower cover plate and the isolation switch to deform, and the signal cannot be collected normally after stepping on the foot switch, making the training equipment function invalid; there are many foot switches, and the charging workload is large.

[0006] In summary, it is necessary to provide a vehicle inspection step recognition method based on a new principle to overcome the above-mentioned technical problems existing in the prior art. Summary of the invention

[0007] In view of the above defects or deficiencies in the prior art, the present invention provides a vehicle inspection pace recognition method and system based on a single-line laser radar, which can fully or partially solve the above technical problems.

[0008] One aspect of the present invention provides a method for identifying vehicle inspection steps based on a single-line laser radar, comprising the following steps:

[0009] Respond to the scanning operation of the single-line laser radar on the vehicle inspection training site and obtain the point cloud data of each scan;

[0010] Query each standard step position in the step position list, and generate a circular area with a preset radius with each standard step position as the center;

[0011] Determine whether the number of data points in the point cloud data scanned more than twice in a row that fall into a certain circular area exceeds a predetermined number, if yes, the standard step position corresponding to the circular area is considered to be in a stepped state, otherwise it is considered to be in a not stepped state;

[0012] The state change of each standard step position is compared with the step position and step sequence specified in the vehicle inspection method, and the recognition result of the inspection step is output.

[0013] Furthermore, the inspection pace calibration step is also included:

[0014] Respond to the scanning operation of the single-line laser radar on the inspection vehicle step calibration site to obtain the point cloud data of this scan;

[0015] Extract the point cloud data whose minimum spacing with the data points of the background point cloud is greater than the first predetermined distance from the point cloud data of this scan, use the extracted point cloud data as the non-background point cloud, cluster the points in the non-background point cloud whose spacing is less than the second predetermined distance into the target point cloud, and retain the two target point clouds with the largest number of points;

[0016] Calculate the centroid of each target point cloud obtained in this scan, search the historical target point cloud obtained in the last scan in the historical target point cloud list, and determine whether the distance between the centroid of the target point cloud and the centroid of the historical target point cloud obtained in the last scan is less than a third predetermined distance. If so, update the most recent tracking time of the historical target point cloud to the generation time of the target point cloud obtained in this scan; otherwise, add the target point cloud obtained in this scan as a new historical target point cloud to the historical target point cloud list, delete the previous historical target point cloud saved in the historical target point cloud list, and record the generation time of the new historical target point cloud;

[0017] The time difference between the most recent tracking time and the generation time of the historical target point cloud in the historical target point cloud list is calculated. If it exceeds a preset threshold, the centroid coordinates of the historical target point cloud are used as the standard step position, and the standard step position is saved in the step position list.

[0018] Furthermore, the step of saving the standard step position to the step position list further includes:

[0019] Generating a marking point corresponding to the standard step position on a display screen of the control unit;

[0020] In response to the operator's input operation, the marking points that do not match the provisions of the vehicle inspection law are deleted on the display screen, and the standard step positions corresponding to the remaining marking points are saved in the step position list.

[0021] Furthermore, before the inspection pace calibration step, a background point cloud generation step is also included:

[0022] In response to the scanning operation of the single-line laser radar on the scene background, the accumulated point cloud data is obtained to form a gridded background point cloud.

[0023] Furthermore, the first predetermined distance and the second predetermined distance are twice the grid size of the background point cloud; and the third predetermined distance is the grid size of the background point cloud.

[0024] Another aspect of the present invention further provides a vehicle inspection step recognition system based on a single-line laser radar, comprising:

[0025] The scanning data acquisition module is configured to respond to the scanning operation of the single-line laser radar on the vehicle inspection step training site and acquire point cloud data for each scan;

[0026] A position generating module is configured to query each standard step position in the step position list, and generate a circular area with a preset radius with each standard step position as the center;

[0027] A judgment module is configured to judge whether the number of data points of the point cloud data scanned more than twice in a row that fall into a certain circular area exceeds a predetermined number, and if so, the standard step position corresponding to the circular area is determined as being stepped on, otherwise it is determined as not being stepped on;

[0028] The output module is configured to compare the state change of each standard step position with the step position and step sequence specified in the vehicle inspection method, and output the recognition result of the vehicle inspection step.

[0029] Furthermore, it also includes a vehicle inspection step calibration module, which is configured to:

[0030] Respond to the scanning operation of the single-line laser radar on the inspection vehicle step calibration site to obtain the point cloud data of this scan;

[0031] Extract the point cloud data whose minimum spacing with the data points of the background point cloud is greater than the first predetermined distance from the point cloud data of this scan, use the extracted point cloud data as the non-background point cloud, cluster the points in the non-background point cloud whose spacing is less than the second predetermined distance into the target point cloud, and retain the two target point clouds with the largest number of points;

[0032] Calculate the centroid of each target point cloud obtained in this scan, search the historical target point cloud obtained in the last scan in the historical target point cloud list, and determine whether the distance between the centroid of the target point cloud and the centroid of the historical target point cloud obtained in the last scan is less than a third predetermined distance. If so, update the most recent tracking time of the historical target point cloud to the generation time of the target point cloud obtained in this scan; otherwise, add the target point cloud obtained in this scan as a new historical target point cloud to the historical target point cloud list, delete the previous historical target point cloud saved in the historical target point cloud list, and record the generation time of the new historical target point cloud;

[0033] The time difference between the most recent tracking time and the generation time of the historical target point cloud in the historical target point cloud list is calculated. If it exceeds a preset threshold, the centroid coordinates of the historical target point cloud are used as the standard step position, and the standard step position is saved in the step position list.

[0034] Furthermore, the vehicle inspection pace calibration module is further configured as follows:

[0035] Generating a marking point corresponding to the standard step position on a display screen of the control unit;

[0036] In response to the operator's input operation, the marking points that do not match the provisions of the vehicle inspection law are deleted on the display screen, and the standard step positions corresponding to the remaining marking points are saved in the step position list.

[0037] Furthermore, it also includes a background point cloud generation module, which is configured as follows:

[0038] In response to the scanning operation of the single-line laser radar on the scene background, the accumulated point cloud data is obtained to form a gridded background point cloud.

[0039] Furthermore, the first predetermined distance and the second predetermined distance are twice the grid size of the background point cloud; and the third predetermined distance is the grid size of the background point cloud.

[0040] The vehicle inspection step recognition method and system based on single-line laser radar provided by the present invention have the following beneficial effects:

[0041] (1) New algorithm. This paper provides a non-contact vehicle inspection step recognition algorithm that can identify both the position of the step and the sequence of the step. The algorithm is accurate and reliable.

[0042] (2) Universality. It supports all existing locomotive and vehicle inspection methods and new inspection methods for new locomotive and vehicle models. It does not require a dedicated training ground and can be used for a long time after one investment.

[0043] (3) Portability. The number of devices required is small, and the devices communicate wirelessly. If training is required in different workshops, the training department can just transport the equipment to that workshop, and there is no need for inspectors to be trained in other places.

[0044] (4) Easy to deploy. No engineering construction is required, just place the step scanning device near the locomotive or vehicle;

[0045] (5) Easy to use. The placement of the step scanning device does not occupy the inspector's workspace. The training scenario is consistent with the actual application scenario. The inspector will not damage the step scanning device during the training process, and the step scanning device will not affect the inspector's inspection practice.

[0046] (6) Long service life. During the training process, there is no contact between the equipment and the inspector, which reduces the possibility of human damage;

[0047] (7) Easy maintenance. Inspection training only requires 2 to 3 step scanning devices. Since the step scanning device is small in size and portable, there is no contact with the inspector during training, so the daily maintenance workload is very small. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0049] Figure 1 It is a schematic diagram of the 25-step inspection method for the speed-up railway freight car C64K (H) in the prior art;

[0050] Figure 2 It is a real installation diagram of a fixed mechanical foot switch in the prior art;

[0051] Figure 3 It is a schematic diagram of the system structure of a movable foot switch in the prior art;

[0052] Figure 4 It is a system structure diagram of a locomotive maintenance training device based on a single-line laser radar provided by an embodiment of the present application;

[0053] Figure 5 This is a schematic diagram of a scenario of a locomotive maintenance training device based on a single-line laser radar provided by an embodiment of the present application;

[0054] Figure 6 is a point cloud data diagram of a single-line laser radar provided by an embodiment of the present application;

[0055] Figure 7 It is a flowchart of a vehicle inspection step recognition method based on a single-line laser radar provided by an embodiment of the present application;

[0056] Figure 8 It is a schematic diagram of a step calibration process provided by an embodiment of the present application;

[0057] Fig. 9 is a schematic diagram of a step recognition process provided by an embodiment of the present application;

[0058] Fig.10 It is a structural schematic diagram of a vehicle inspection step recognition system based on a single-line laser radar provided by an embodiment of the present application;

[0059] Fig.11 It is a structural diagram of a vehicle inspection step recognition module provided by an embodiment of the present application. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "said" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.

[0062] It should be understood that although the terms first, second, third, etc. may be used to describe the acquisition modules in the embodiments of the present invention, the acquisition modules should not be limited to these terms. These terms are only used to distinguish the acquisition modules from each other.

[0063] The word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)", depending on the context.

[0064] It should be noted that the directional words such as "upper", "lower", "left", and "right" described in the embodiments of the present invention are described at the angles shown in the drawings and should not be understood as limiting the embodiments of the present invention. In addition, in the context, it should also be understood that when it is mentioned that an element is formed "on" or "under" another element, it can not only be formed directly "on" or "under" another element, but also be formed "on" or "under" another element indirectly through an intermediate element.

[0065] See also Figure 4 A locomotive maintenance training device 100 based on a single-line laser radar in this embodiment includes: a step scanning unit 101, a check point perception unit 106, a video acquisition unit 102, a communication control unit, a voice prompt unit 107, an audio acquisition unit 108 and a step position indication unit, wherein the communication control unit is communicatively connected with the step scanning unit 101, the check point perception unit 106, the video acquisition unit 102, the voice prompt unit 107, and the audio acquisition unit 108.

[0066] Figure 5 1 is a schematic diagram of a scenario of a locomotive maintenance training device 100 based on a single-line laser radar according to this embodiment.

[0067] Specifically, the locomotive maintenance training device 100 of this embodiment includes:

[0068] The step scanning unit 101 includes a single-line laser radar, a wireless bridge (or a WIFI module) and a power supply, and the wireless bridge is connected to the communication unit. Among them, the single-line laser radar uses a single-beam laser transmitter to rotate at a constant speed inside the radar, and emits a laser once every rotation angle. After polling a certain angle, a frame of complete data is formed, so the data of the single-line laser radar can be regarded as a row of dot matrix at the same height. The scanning plane of the single-line laser radar in this embodiment is parallel to the ground and is set at a predetermined height from the ground, which is used to obtain the position data of the step. Preferably, the radar scanning surface is parallel to the ground and is about 15 cm above the ground, which has the best effect. More preferably, the step scanning unit 101 is a box-type device or an integrated device, which is set at the front side of the trainer's direction of travel. The radar scanning surface should be able to cover all step positions. If there is an obstruction, multiple step scanning units 101 can be placed, and the radar scanning surfaces of multiple step scanning devices should be able to cover all step positions. Figure 6 It is a point cloud data diagram of the step position obtained by the single-line laser radar of this embodiment.

[0069] Furthermore, the training device may also include a photoelectric detector 109 for detecting whether the scanning plane of the single-line laser radar is horizontal. Specifically, the photoelectric detector 109 is arranged at different positions around the step scanning unit 101 and is arranged facing the step scanning unit 101. The photoelectric detector 109 includes a number of laser receivers arranged horizontally, and each laser receiver is provided with an indicator light. The laser receiver is used to receive the laser beam emitted by the single-line laser radar and light up the corresponding indicator light after successful reception. When the user adjusts the laser scanning plane of the step scanning unit 101, the user can judge whether the current scanning plane of the single-line laser radar is horizontal according to the number of indicator lights that light up.

[0070] The inspection sensing unit 106 includes an information button and an inspection hammer adsorbed on the part to be inspected on the locomotive body. The information button is provided with an adsorption device and a radio frequency tag (or radio frequency card). The inspection hammer is wirelessly connected with the radio frequency tag and the communication unit. The inspection hammer is used to obtain the number information in the radio frequency tag and send it to the communication unit. The control unit identifies the number information received by the communication unit to confirm whether the inspector's inspection hammer hits the predetermined inspection position of the locomotive. Preferably, the adsorption device of the information button adopts a magnetic adsorption element or a vacuum adsorption element, and the radio frequency tag is an RFID tag or an NFC tag. The inspection hammer is provided with an RF signal transmitter and an RF signal receiver corresponding to the type of RF tag. When the inspection hammer hits the information button, the RF signal transmitter of the inspection hammer sends radio waves to the RF tag or RF card. The RF tag or RF card receives the RF signal and generates current through electromagnetic induction. The current drives the RF tag or RF card to send the number information in the card. The RF signal receiver receives the RF signal with the number information, and then sends it to the communication unit through the number information transmission module. The control unit can identify whether the inspector hits the information button at the predetermined position of the locomotive by identifying the number information, thereby judging the training effect of the inspector.

[0071] The video acquisition unit 102 includes a camera, a pan-tilt head and a WIFI module (or a wireless bridge). The camera is arranged on the pan-tilt head, and the WIFI module (or a wireless bridge) is wirelessly connected to the communication unit.

[0072] The audio collection unit 108 is preferably a wireless headset. Specifically, during the assessment phase of the training, the trainees are required to wear wireless headsets to collect the information data of the trainees' oral inspection and store the audio data. The assessment teacher plays back the audio data through the control unit to check whether the "oral information" of the trainees during the inspection is standard and whether the information button is checked in place.

[0073] The communication control device includes a control unit 104 and a communication unit 105. The communication unit 105 includes a component for wireless communication with the wireless bridge, the WIFI module, the inspection hammer, the voice prompt unit 107, and the audio acquisition unit 108, and is used to obtain the collected data and transmit the control instructions issued by the control unit 104. The control unit 104 is used to collect data through the communication unit 105, determine whether the acquired step action information is correct, determine whether the inspection action of the inspected part is complete, and the collected data and the judgment results are recorded in the database and displayed on the interface. The control unit can be a PDA, a laptop, a PC or a smart phone, etc.

[0074] The voice prompt unit 107 is preferably a Bluetooth speaker or wireless headset, so as to provide voice prompt guidance during training and provide voice prompts for incorrect actions. The control unit can directly broadcast the trainer's voice through the voice prompt unit 107, or convert the broadcast text into sound data and transmit it to the voice prompt unit 107 through the headphone jack line of the control unit.

[0075] The step position indicating unit is a step position indicating sticker 103 pasted on the ground, or a step indicating blanket laid on the ground, so as to facilitate the inspector to practice the step action according to the step indicating mark.

[0076] The working principle of the locomotive maintenance training device based on single-line laser radar in this embodiment is as follows: the inspection method stipulates the step position trajectory and the parts to be inspected at each step, and the step scanning unit collects radar point cloud to detect whether the step action is correct; the inspection point perception unit is used to detect whether the parts to be inspected are fully inspected; during the training process, guidance is given through voice, and voice prompts are given for wrong steps, missed steps or missed parts to be inspected so that the inspector can correct them; during the assessment process, no voice guidance is given, and after the inspector has completed the operation, points are deducted for wrong steps, missed steps, and missed parts, and the accuracy and proficiency of the inspector in mastering the inspection method are determined by the test scores.

[0077] Another embodiment of the present invention provides a method for identifying vehicle inspection steps based on a single-line laser radar, which can identify vehicle inspection steps without contacting an actual detection device. Figure 7 , the method comprises the following steps:

[0078] Step S101, in response to the scanning operation of the single-line laser radar on the scene background, the accumulated point cloud data is obtained to form a gridded background point cloud.

[0079] Specifically, before formally conducting pace training for inspectors, it is necessary to calibrate the standard pace position that complies with the inspection law, which includes both the calibration of the on-site ground position and the calibration of the single-line laser radar gridded point cloud data position. Since the calibration of the pace position requires identifying the point cloud positions of the two feet of the on-site trainees on the laser radar, before calibration, it is necessary to first perform a single-line laser radar scan on the pure background when there are no trainees at the training site to obtain the pure on-site background point cloud data after excluding the personnel. These background point cloud data include radar point cloud data of objects such as locomotives, walls, and ground obstacles at the site, which can outline the approximate position distribution of the data points corresponding to the on-site background in the gridded background point cloud.

[0080] For example: the main control software of the control unit starts the point cloud accumulation of the single-line laser radar. The single-line scanning radar scans the entire scene 25 times per second, and each scan generates 1080 scanning points. After accumulating for 10 seconds, the point cloud accumulation stops, and the accumulated point cloud is gridded and downsampled to form a grid background point cloud. The grid size of this embodiment can be determined according to the radar angular resolution and scanning range. After the grid background is generated, the voice prompt unit plays the voice prompt "Background learning is complete, please calibrate". After hearing the voice, the training teacher can calibrate the steps.

[0081] Step S102, calibration of vehicle inspection steps.

[0082] Specifically, the calibration personnel hold chalk or a marker, walk along the route required by the vehicle inspection method, stay at the specified step position for more than 3 seconds, and pass the non-specified step position within 3 seconds. Use chalk or a marker to mark the specified position so that the step indicator sticker can be affixed for the inspection personnel to refer to the landing position during training. Figure 8 , while the calibration personnel are walking, the main control software receives the scanning point cloud of the single-line laser radar in real time, and each scanning point cloud is processed as follows:

[0083] Step S1021, in response to the scanning operation of the single-line laser radar on the inspection step calibration site, obtain the point cloud data of this scanning.

[0084] Specifically, a single scan point cloud data of a single-line laser radar is received. This embodiment is configured by default to upload point cloud data once every 40 milliseconds, and each point cloud contains 1080 data points. In addition to the background point cloud obtained in this scan point cloud data, the point cloud data of the two feet of the calibration personnel on site are also obtained. These point cloud data are called non-background point clouds, but the background point clouds and non-background point clouds are mixed together, and the two need to be distinguished to obtain the step position information.

[0085] Step S1022, extract the point cloud data whose minimum spacing with the data points of the background point cloud is greater than the first predetermined distance in the point cloud data of this scan, use the extracted point cloud data as the non-background point cloud, cluster the points in the non-background point cloud whose spacing is less than the second predetermined distance into a target point cloud, and retain the two target point clouds with the largest number of points.

[0086] Specifically, the points in the scanned point cloud data whose minimum distance to the data points of the grid background point cloud is greater than 2 times the grid size are extracted as non-background point clouds. The reason for this is that the point cloud data of the two feet of the calibration personnel and the background point cloud data need to be clearly distinguished, so a first predetermined distance of "2 times the grid size" is set as the distinction threshold. Only the data points with the minimum spacing from the background point cloud greater than this threshold can be used as non-background point clouds, that is, the point cloud data of the two feet of the calibration personnel. Then, the points in the non-background point cloud whose spacing is less than the second predetermined distance are clustered into target point clouds. The reason for this is that the calibration personnel has two feet, and each foot corresponds to a number of point cloud data after a single scan (the grid size of this embodiment is set to 8cm, and each foot may correspond to more than 3 point cloud data after a single scan). In order to distinguish whether the data points in these point cloud data are left by two feet or one foot, this embodiment specifically sets a second predetermined distance of "2 times the grid size" as the distinction threshold, and clusters the points in the non-background point cloud with a spacing less than "2 times the grid size" into a target point cloud, which means that the target point cloud obtained by clustering belongs to the point cloud data of one foot. The two target point clouds with the largest number of points are retained as the target point clouds formed by the two feet, and the other target point clouds are eliminated as invalid data.

[0087] Step S1023, calculate the center of mass of each target point cloud obtained in this scan, search the historical target point cloud obtained in the last scan in the historical target point cloud list, and determine whether the distance between the center of mass of the target point cloud and the center of mass of the historical target point cloud obtained in the last scan is less than a third predetermined distance. If so, update the most recent tracking time of the historical target point cloud to the generation time of the target point cloud obtained in this scan; otherwise, add the target point cloud obtained in this scan as a new historical target point cloud to the historical target point cloud list, delete the previous historical target point cloud saved in the historical target point cloud list, and record the generation time of the new historical target point cloud.

[0088] Specifically, the single-line laser radar can obtain the target point cloud data corresponding to the two feet of the calibration personnel each time it scans, calculate the centroid of each target point cloud obtained in this scan, and then search the historical target point cloud obtained in the last scan from the data list. By comparing the centroid positions of the two, it can be known whether the actual positions of the two feet have changed during the two scans. In order to eliminate invalid actions of the calibration personnel's feet, such as stomping, lifting feet, etc., which cause changes in the centroid of the target point cloud of the two feet, this embodiment sets a third predetermined distance of "1 times the grid size" as a distinguishing threshold. Specifically, if the distance between the centroid of the target point cloud of this scan and the centroid of the historical target point cloud obtained in the last scan is less than "1 times the grid size", it means that the last step position has not changed substantially. At this time, the most recent tracking time of the historical target point cloud is updated to the generation time of the target point cloud obtained in this scan, which means that the step position has not moved substantially from the beginning of generation to the moment of this scan. If the distance between the centroid of the target point cloud of this scan and the centroid of the historical target point cloud obtained in the previous scan is greater than "1 times the grid size", it means that the position of the step has substantially changed during this scan, so the target point cloud obtained in this scan is added to the historical target point cloud list as a new historical target point cloud, the previous historical target point cloud saved in the historical target point cloud list is deleted, and the generation time of the new historical target point cloud is recorded. In other words, the target point cloud obtained in this scan is used as a new target point cloud generated after the step is substantially moved, indicating that a new step has been taken. At this time, the time when the new step is generated should be recorded to facilitate the subsequent judgment of the maintenance time of the new step.

[0089] Step S1024, calculate the time difference between the most recent tracking time and the generation time of the historical target point cloud in the historical target point cloud list. If it exceeds a preset threshold, the centroid coordinates of the historical target point cloud are used as the standard step position, and the standard step position is saved in the step position list.

[0090] Specifically, since the calibration personnel will stay at the standard step position for more than 3 seconds according to regulations, when the historical target point cloud is maintained for more than 3 seconds from the time it is generated to the most recent tracking time, it can be considered that the center of mass of the historical target point cloud is the standard step position. At this time, the standard step position can be saved in the step position list.

[0091] Furthermore, in the actual calibration process, the calibration personnel may stay at a position other than the standard step position specified in the vehicle inspection law for more than 3 seconds for various reasons, which may cause the centroid of the historical target point cloud corresponding to the position to be mistaken as the standard step position, so it is necessary to eliminate such position data. Specifically, a mark point corresponding to the standard step position is generated on the display screen of the control unit, such as a circle, triangle, etc., and in response to the input operation of the operator, the mark points that do not match the provisions of the vehicle inspection law are deleted on the display screen, and the standard step positions corresponding to the remaining mark points are saved in the step position list, and then the remaining mark points are numbered in sequence for reference during the inspection training.

[0092] Step S103, vehicle inspection step recognition.

[0093] See also Fig. 9 During the pace training, the inspector walks and the pace scanning unit scans in real time, and the point cloud of each scan is processed as follows:

[0094] Step S1031, in response to the scanning operation of the single-line laser radar on the vehicle inspection step training site, obtain point cloud data of each scan.

[0095] Step S1032, querying each standard step position in the step position list, and generating a circular area with a preset radius with each standard step position as the center.

[0096] Specifically, due to factors such as the size of the feet and the error of the footfall, the step movement cannot fall precisely on a point, and a tolerance range needs to be set. Therefore, a circle with a radius of about 25 cm is generated for each position element in the step position list. As long as the data point scanned by the radar falls within the range of the circle, there is a possibility of being stepped on.

[0097] Step S1033, determine whether the number of data points in the point cloud data scanned more than twice in a row that fall into a circular area exceeds a predetermined number. If so, the standard step position corresponding to the circular area is identified as being stepped on, otherwise it is identified as not being stepped on.

[0098] Specifically, there are accidental factors in the inspector's footsteps during the step training process, such as stepping, stepping on the spot, and misoperation of the feet, etc. Although the inspector did not actually step on a circular area, the point cloud data scanned by the radar may still fall into the circular area. In order to eliminate invalid data caused by such invalid actions, this embodiment calculates the number of points falling into each circle in the scanned point cloud. If the number of points exceeds the set threshold (default 2), it is considered that the inspector may step on the step position; if there are data points exceeding the set threshold falling into the circle in the scanned point cloud for multiple consecutive times (default 10 times), it is confirmed that the step position corresponding to the circle is stepped on, otherwise it is considered not to be stepped on. When the stepped on state and the not stepped on state change, the state change information of the step position according to the scanning time is output for use by other business modules of the main control software.

[0099] Step S1034, compare the state change of each standard step position with the step position and step sequence specified in the vehicle inspection method, and output the recognition result of the inspection step.

[0100] Specifically, according to the provisions of the standard vehicle inspection method, based on the state change information of the current step position, it can be determined whether the inspector's step position and step sequence are correct, that is, whether a wrong step or a missed step has occurred.

[0101] See also Fig.10 Another embodiment of the present invention further provides a vehicle inspection step recognition system 200 based on a single-line laser radar, comprising a background point cloud generation module 201, a vehicle inspection step calibration module 202 and a vehicle inspection step recognition module 203. The vehicle inspection step recognition system 200 can execute the step recognition method in the above method embodiment.

[0102] Wherein, the background point cloud generation module 201 is configured as follows:

[0103] In response to the scanning operation of the single-line laser radar on the scene background, the accumulated point cloud data is obtained to form a gridded background point cloud.

[0104] The vehicle inspection pace calibration module 202 is configured as follows:

[0105] Respond to the scanning operation of the single-line laser radar on the inspection vehicle step calibration site to obtain the point cloud data of this scan;

[0106] Extract the point cloud data whose minimum spacing with the data points of the background point cloud is greater than the first predetermined distance from the point cloud data of this scan, use the extracted point cloud data as the non-background point cloud, cluster the points in the non-background point cloud whose spacing is less than the second predetermined distance into the target point cloud, and retain the two target point clouds with the largest number of points;

[0107] Calculate the centroid of each target point cloud obtained in this scan, search the historical target point cloud obtained in the last scan in the historical target point cloud list, and determine whether the distance between the centroid of the target point cloud and the centroid of the historical target point cloud obtained in the last scan is less than a third predetermined distance. If so, update the most recent tracking time of the historical target point cloud to the generation time of the target point cloud obtained in this scan; otherwise, add the target point cloud obtained in this scan as a new historical target point cloud to the historical target point cloud list, delete the previous historical target point cloud saved in the historical target point cloud list, and record the generation time of the new historical target point cloud;

[0108] The time difference between the most recent tracking time and the generation time of the historical target point cloud in the historical target point cloud list is calculated. If it exceeds a preset threshold, the centroid coordinates of the historical target point cloud are used as the standard step position, and the standard step position is saved in the step position list.

[0109] Furthermore, the vehicle inspection pace calibration module 202 is configured as follows:

[0110] Generating a marking point corresponding to the standard step position on a display screen of the control unit;

[0111] In response to the operator's input operation, the marking points that do not match the provisions of the vehicle inspection law are deleted on the display screen, and the standard step positions corresponding to the remaining marking points are saved in the step position list.

[0112] See also Fig.11 The vehicle inspection step recognition module 203 includes:

[0113] The scanning data acquisition module 2031 is configured to respond to the scanning operation of the single-line laser radar on the vehicle inspection step training site and acquire the point cloud data of each scan;

[0114] The position generating module 2032 is configured to query each standard step position in the step position list, and generate a circular area with a preset radius with each standard step position as the center;

[0115] The judging module 2033 is configured to judge whether the number of data points in the point cloud data scanned more than twice in a row that fall into a certain circular area exceeds a predetermined number, and if so, the standard step position corresponding to the circular area is determined as being stepped on, otherwise it is determined as not being stepped on;

[0116] The output module 2034 is configured to compare the state change of each standard step position with the step position and step sequence specified in the vehicle inspection method, and output the recognition result of the vehicle inspection step.

[0117] It should be noted that the vehicle inspection pace recognition system 200 based on single-line laser radar provided in this embodiment corresponds to a technical solution that can be used to execute various method embodiments, and its implementation principle and technical effects are similar to the methods, which will not be repeated here.

[0118] The above description is only a preferred embodiment of the present invention. Those skilled in the art should understand that the disclosure scope involved in the present invention is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present invention (but not limited to) by each other to form a technical solution.

Claims

1. A vehicle inspection step recognition method based on a single-line laser radar, characterized in that: The steps include: Responding to the scanning operation of the single-line laser radar on the inspection vehicle pace training site, obtaining the point cloud data of each scan, including: responding to the scanning operation of the single-line laser radar on the inspection vehicle pace calibration site, obtaining the point cloud data of this scan; extracting the point cloud data of this scan whose minimum spacing with the data points of the background point cloud is greater than a first predetermined distance, using the extracted point cloud data as non-background point cloud, clustering the points in the non-background point cloud whose spacing is less than a second predetermined distance into target point clouds, and retaining the two target point clouds with the largest number of points; calculating the centroid of each target point cloud obtained in this scan, searching the historical target point cloud obtained by the last scan in the historical target point cloud list, and judging the centroid of the target point cloud and the centroid of the target point cloud obtained by the last scan. whether the distance between the centroids of the historical target point clouds is less than a third predetermined distance, if so, the most recent tracking time of the historical target point cloud is updated to the generation time of the target point cloud obtained by this scan, otherwise, the target point cloud obtained by this scan is added to the historical target point cloud list as a new historical target point cloud, the previous historical target point cloud saved in the historical target point cloud list is deleted, and the generation time of the new historical target point cloud is recorded; the time difference between the most recent tracking time and the generation time of the historical target point cloud in the historical target point cloud list is calculated, and if it exceeds a preset threshold, the centroid coordinates of the historical target point cloud are used as the standard step position, and the standard step position is saved in the step position list; Query each standard step position in the step position list, and generate a circular area with a preset radius with each standard step position as the center; Determine whether the number of data points in the point cloud data scanned more than twice in a row that fall into a certain circular area exceeds a predetermined number, if yes, the standard step position corresponding to the circular area is considered to be in a stepped state, otherwise it is considered to be in a not stepped state; The state change of each standard step position is compared with the step position and step sequence specified in the vehicle inspection method, and the recognition result of the inspection step is output.

2. The vehicle inspection step recognition method based on single-line laser radar according to claim 1 is characterized in that: The step of saving the standard step position to the step position list also includes: Generating a marking point corresponding to the standard step position on a display screen of the control unit; In response to the operator's input operation, the marking points that do not match the provisions of the vehicle inspection law are deleted on the display screen, and the standard step positions corresponding to the remaining marking points are saved in the step position list.

3. The vehicle inspection step recognition method based on single-line laser radar according to claim 1 is characterized in that: Before the inspection pace calibration step, the background point cloud generation step is also included: In response to the scanning operation of the single-line laser radar on the scene background, the accumulated point cloud data is obtained to form a gridded background point cloud.

4. The vehicle inspection step recognition method based on single-line laser radar according to claim 1 is characterized in that: The first predetermined distance and the second predetermined distance are twice the grid size of the background point cloud; The third predetermined distance is the grid size of the background point cloud.

5. A vehicle inspection step recognition system based on a single-line laser radar, characterized in that: include: The scanning data acquisition module is configured to respond to the scanning operation of the single-line laser radar on the inspection vehicle pace training site and obtain the point cloud data of each scan, including: responding to the scanning operation of the single-line laser radar on the inspection vehicle pace calibration site to obtain the point cloud data of this scan; extracting the point cloud data of this scan whose minimum spacing with the data points of the background point cloud is greater than a first predetermined distance, using the extracted point cloud data as non-background point cloud, clustering the points in the non-background point cloud whose spacing is less than a second predetermined distance into target point cloud, and retaining the two target point clouds with the largest number of points; calculating the centroid of each target point cloud obtained in this scan, searching the historical target point cloud list for the historical target point cloud obtained by the last scan, and determining the centroid of the target point cloud Whether the distance between the centroid of the historical target point cloud obtained by the last scan is less than a third predetermined distance, if so, the most recent tracking time of the historical target point cloud is updated to the generation time of the target point cloud obtained by this scan, otherwise, the target point cloud obtained by this scan is added to the historical target point cloud list as a new historical target point cloud, the previous historical target point cloud saved in the historical target point cloud list is deleted, and the generation time of the new historical target point cloud is recorded; the time difference between the most recent tracking time and the generation time of the historical target point cloud in the historical target point cloud list is calculated, and if it exceeds a preset threshold, the centroid coordinates of the historical target point cloud are used as the standard step position, and the standard step position is saved in the step position list; A position generating module is configured to query each standard step position in the step position list, and generate a circular area with a preset radius with each standard step position as the center; A judgment module is configured to judge whether the number of data points of the point cloud data scanned more than twice in a row that fall into a certain circular area exceeds a predetermined number, and if so, the standard step position corresponding to the circular area is determined as being stepped on, otherwise it is determined as not being stepped on; The output module is configured to compare the state change of each standard step position with the step position and step sequence specified in the vehicle inspection method, and output the recognition result of the vehicle inspection step.

6. The vehicle inspection step recognition system based on single-line laser radar according to claim 5 is characterized in that: The vehicle inspection step calibration module is further configured as follows: Generating a marking point corresponding to the standard step position on a display screen of the control unit; In response to the operator's input operation, the marking points that do not match the provisions of the vehicle inspection law are deleted on the display screen, and the standard step positions corresponding to the remaining marking points are saved in the step position list.

7. The vehicle inspection step recognition system based on single-line laser radar according to claim 5 is characterized in that: It also includes a background point cloud generation module, which is configured as follows: In response to the scanning operation of the single-line laser radar on the scene background, the accumulated point cloud data is obtained to form a gridded background point cloud.

8. The vehicle inspection step recognition system based on single-line laser radar according to claim 5 is characterized in that: The first predetermined distance and the second predetermined distance are twice the grid size of the background point cloud; The third predetermined distance is the grid size of the background point cloud.

Images