A method for detecting obstacles in the gap of shielded doors based on edge computing laser radar

Abstract

The present invention relates to the field of rail transit, and in particular to a method for detecting obstacles in the gap of a shielded door based on an edge computing laser radar, which is applied to a laser radar based on ToF, wherein the laser radar is provided with an edge computing module, and the method comprises: scanning the environment by the laser radar to collect current detection data; performing data processing on the current detection data by the edge computing module to obtain the current detection boundary; judging whether the current detection boundary intrudes into the defense zone boundary; if so, it indicates that there is an obstacle in the defense zone, and the obstacle alarm information is uploaded; if not, it indicates that there is no obstacle in the defense zone; wherein the data processing sequentially comprises straight-through filtering processing, conditional filtering processing, data conversion processing and statistical filtering processing, so as to filter background interference data. The scheme of the present invention can achieve full coverage of the gap between the shielded door and the car door, has strong anti-interference ability, and has a detection accuracy of 10-20mm, and can accurately locate obstacles.

Description

Technical Field

[0001] The present invention relates to the field of rail transportation, and in particular to a method for detecting obstacles in the gap of a shielding door based on an edge computing laser radar. Background Art

[0002] Due to the objective existence of the gap between the platform door and the train, there is a risk of passengers being trapped between the platform door and the train body when getting on and off the train, resulting in the risk of personal injury. As new subway projects in cities such as Beijing, Shanghai, and Guangzhou have successively adopted fully automatic unmanned driving modes, the new operating scenarios have put forward higher requirements for the safety protection of the gap between the platform door and the train.

[0003] Traditional gap detection methods include physical baffles, manual lookout light strips, and photoelectric cross-beam detection. Physical baffles are easy to loosen and invade during installation, and can only detect larger objects within 500mm from the ground. They do not have full-area detection capabilities and detection accuracy. Manual lookout light strips rely on human judgment and are subjective. They are not suitable for curved platforms or fully automatic driving lines, and cannot be linked to other systems. Photoelectric cross-beam detection is not suitable for curved platforms. It is prone to false alarms due to interference from strong light, vibration, dust and iron filings, and has no alarm door positioning function. Since traditional gap detection methods have corresponding shortcomings and limitations and cannot meet the higher gap safety protection requirements in new operating scenarios, how to provide a shield door gap obstacle detection method that can meet higher safety protection requirements has become a technical problem that needs to be solved urgently in this field. Summary of the invention

[0004] The object of the present invention is to provide a method for detecting obstacles in the gap of a shielding door based on edge computing laser radar, so as to solve the above-mentioned problems existing in the prior art.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A method for detecting obstacles in the gap of a shielding door based on an edge computing laser radar is applied to a ToF-based laser radar, wherein the laser radar is provided with an edge computing module, and the method comprises:

[0007] Scanning the environment through the laser radar to collect current detection data;

[0008] Processing the current detection data by the edge computing module to obtain a current detection boundary;

[0009] Determine whether the current detection boundary intrudes into the defense zone boundary; if so, it indicates that there is an obstacle in the defense zone, and the obstacle alarm information is uploaded; if not, it indicates that there is no obstacle in the defense zone;

[0010] The data processing includes through-filtering, conditional filtering, data conversion and statistical filtering in sequence, so as to filter the background interference data.

[0011] Preferably, the data processing specifically includes:

[0012] By presetting the detection angle, the data exceeding the preset detection angle is filtered;

[0013] By presetting the detection radius, the data exceeding the preset detection radius is filtered;

[0014] Convert the filtered valid data into two-dimensional polar coordinates of detection distance and detection angle and store them;

[0015] Based on the two-dimensional polar coordinates, the average detection distance of each data point in multiple detection cycles and the change amount relative to adjacent data points are statistically analyzed, and the data points whose average detection distance exceeds the statistical mean and variance preset thresholds and the data points whose change amount exceeds the standard deviation preset threshold are filtered.

[0016] Preferably, the defense zone boundary is obtained by the following method:

[0017] Scanning an obstacle-free environment by using the laser radar to collect original detection data;

[0018] Performing the data processing on the original detection data by the edge computing module to obtain a detection boundary;

[0019] Based on the data coordinates of the detection boundary, the defense zone boundary is obtained by shrinking the preset safety distance toward the data origin.

[0020] Preferably, the laser radar is also provided with a wireless communication device, which is connected to a mobile terminal through the wireless communication device, and the detection results are displayed and parameters of the laser radar are set through the mobile terminal.

[0021] Furthermore, the laser radar is also provided with a temperature sensor and a posture sensor; before the laser radar scans, a status self-inspection is performed, and when any inspection result does not meet the preset standard, a maintenance prompt is issued; the inspection items of the status self-inspection include: the temperature measured by the temperature sensor, the three-dimensional posture information measured by the posture sensor, the communication status and the storage status.

[0022] Furthermore, the preset standard of the three-dimensional posture information is the three-dimensional posture information when the laser radar obtains the defense zone boundary, and the three-dimensional posture information includes pitch angle, left and right yaw angle and vertical rotation angle.

[0023] Preferably, before determining whether the current detection boundary intrudes into the defense zone boundary, a dust cover detection is also performed, including:

[0024] Determine a data point within a first preset distance from the data origin as a dust point, and issue a dust prompt if the dust point exists;

[0025] Determine a data point between the first preset distance and the second preset distance from the data origin as an occlusion point, and issue an occlusion prompt if the occlusion point exists;

[0026] Wherein, the second preset distance is greater than the first preset distance.

[0027] Further, determining whether the current detection boundary intrudes into the defense zone boundary specifically includes:

[0028] The data coordinates of the current detection boundary are located within the defense zone boundary and the data point outside the second preset distance from the data origin is determined as an obstacle point. If the obstacle point exists, it means that the current detection boundary invades the defense zone boundary.

[0029] Furthermore, if the current detection boundary intrudes into the defense zone boundary, it also includes obtaining the shape, size and position of the corresponding obstacle based on the data information of the obstacle point, and performing clustering restoration to identify the corresponding risk level and generate obstacle alarm information.

[0030] The beneficial effects of the present invention are:

[0031] The present invention adopts a ToF-based laser radar, which is equipped with an edge computing module. The acquired data is processed at the front-end edge computing to save back-end computing power. When the train door and the platform shield door are closed, the vertical surface of the narrow space between the train door and the shield door can be scanned. When there is an obstacle in the defense zone, the round-trip time of the laser beam will change, so that the obstacle can be detected and located. It can effectively detect obstacles with a diameter of 20mm (backpack straps, thin ropes, etc.), with a maximum detection accuracy of 10mm, and has the characteristics of accurate positioning and strong anti-interference. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flow chart of a method for detecting obstacles in a gap of a shielding door based on edge computing laser radar in an embodiment of the present invention;

[0033] Figure 2 is a schematic diagram of a data processing flow in an embodiment of the present invention;

[0034] Figure 3 is a schematic diagram of straight-through filtering in an embodiment of the present invention;

[0035] Figure 4is a schematic diagram of conditional filtering in an embodiment of the present invention;

[0036] Figure 5 Schematic diagram of the data range after filtering in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation methods described herein are only used to explain the present invention and are not used to limit the present invention.

[0038] like Figure 1 As shown, the present invention provides a shielding door gap obstacle detection method based on edge computing laser radar, which is applied to a ToF-based laser radar, wherein the laser radar is provided with an edge computing module, and the method comprises:

[0039] S101, scanning the environment by using the laser radar to collect current detection data.

[0040] S102: Process the current detection data by the edge computing module to obtain a current detection boundary.

[0041] S103, determining whether the current detection boundary intrudes into the defense zone boundary; if so, it indicates that there is an obstacle in the defense zone, and the obstacle alarm information is uploaded; if not, it indicates that there is no obstacle in the defense zone.

[0042] The data processing includes through-filtering, conditional filtering, data conversion and statistical filtering in sequence, so as to filter the background interference data.

[0043] In this embodiment, the laser radar is integrated with a transmitter and receiver, and uses a 905nm wavelength light beam, which is most suitable for propagation in the air. The detection radius of a single laser radar is 7000mm and the detection angle is 100°. Each set of laser radars can cover the corresponding sliding doors and fixed doors, and has a self-learning function for the detection area defense zone, and can set the detection area of ​​any contour. The use of edge computing technology does not require the radar data to be transmitted to the host computer, and the detection results can be directly judged. Different vehicle conditions (full load, no load) and different platform gaps (straight platform, curved platform) have no effect on the detection light curtain, and can effectively detect obstacles with a diameter of 20mm (backpack straps, thin ropes, etc.), with a detection accuracy of 10-20mm.

[0044] In this embodiment, the edge computing module is responsible for filtering, converting, and filtering background interference data from the laser detector; removing unstable point data and stabilizing frequently jumping data. Figure 2 As shown, the data processing specifically includes:

[0045] S201, filtering data exceeding the preset detection angle by using the preset detection angle.

[0046] S202: filtering data exceeding the preset detection radius by using the preset detection radius.

[0047] S203, converting the filtered valid data into two-dimensional polar coordinates of detection distance and detection angle and storing them.

[0048] S204, based on the two-dimensional polar coordinates, statistically analyze the average detection distance of each data point in multiple detection cycles and the change amount relative to adjacent data points, and filter out the data points whose average detection distance exceeds the statistical mean and variance preset thresholds and the data points whose change amount exceeds the standard deviation preset threshold.

[0049] In this embodiment, the filtering of the laser radar source data is divided into two steps, such as Figure 3 As shown, the first step adopts the straight-through filtering method, by defining the detection start and end angles α-β, filtering the data exceeding the start and end angle range, and only passing the detection data within the α-β angle range backward.

[0050] like Figure 4 As shown, the second step adopts the conditional filtering method, by defining the detection radius L, the data exceeding the preset detection radius is filtered, and the specific operation is to filter the data whose flight time exceeds 2L / c time (c: speed of light).

[0051] like Figure 5 As shown, after two-step filtering, the effective data range is reduced, effectively reducing the amount of calculation in the next step.

[0052] The filtered valid data is calculated based on the flight time measurement and converted into the corresponding detection distance (the distance from the data point to the laser radar), and the calculation formula is L1 = δt*c / 2 (δt: laser pulse round-trip time, c: speed of light, L1: target distance of the first measuring point). The detection distance is combined with the detection angle to form a two-dimensional polar coordinate, and all valid coordinates of a detection cycle are stored.

[0053] The storage data form is shown in Table 1.

[0054] Table 1

[0055] Serial number angle distance Light intensity 1 α1 L1 l1 2 α2 L2 l2 …

[0056] The light intensity corresponding to the data point and the data point coordinates may also be saved simultaneously, and the background environment interference measurement points with light intensity less than a set value (such as 3%) may be filtered.

[0057] Finally, through statistical filtering, the average distance of each measuring point in multiple cycles and the change of adjacent measuring points are statistically analyzed. Points whose average distance is outside the statistical mean and variance standards are judged as unstable points, and points with large standard deviations of the change of adjacent measuring points and the above-mentioned unstable points are removed.

[0058] In this embodiment, when the laser radar is in a normal working environment, it waits for the self-learning instruction provided by the host computer to learn the current detection boundary of the device to form a defense zone. Specifically, the defense zone boundary is obtained by the following method:

[0059] The obstacle-free environment is scanned by the laser radar to collect original detection data.

[0060] The edge computing module performs the data processing on the original detection data to obtain a detection boundary.

[0061] Based on the data coordinates of the detection boundary, the defense zone boundary is obtained by shrinking the preset safety distance toward the data origin.

[0062] In this embodiment, the laser radar can be connected through a redundant CAN bus to support linkage with signals, platform doors, and video systems. A wireless communication device can also be provided to connect a mobile terminal through the wireless communication device, and the detection results can be displayed and parameters of the laser radar can be set through the mobile terminal. Specifically, by connecting the laser radar hotspot through a handheld tablet, a dedicated APP can be used to set parameters such as alarm sensitivity, detection distance, and radar opening angle. The self-learned defense zone boundary of the laser radar and the detection graphics of the current detection boundary can be intuitively viewed on the handheld tablet.

[0063] In this embodiment, the laser radar is also provided with a temperature sensor and a posture sensor; before the laser radar scans, it also performs a self-check of the status, and when any of the check results do not meet the preset standard, a maintenance prompt is issued; the check items of the self-check of the status include: the temperature measured by the temperature sensor, the three-dimensional posture information measured by the posture sensor, the communication status and the storage status. Among them, the preset standard of the three-dimensional posture information is the three-dimensional posture information when the laser radar obtains the defense zone boundary, and the three-dimensional posture information includes the pitch angle, the roll left and right angle and the yaw vertical direction angle.

[0064] In this embodiment, the laser radar starts to be powered on and performs equipment self-test, and detects the working status of equipment components, temperature, storage space preset data, communication status and other modules. If the gap detection radar finds that the components are unresponsive, the temperature is abnormal, the storage space preset data is wrong, the communication is interrupted, etc. during the self-test, the equipment fault status is comprehensively judged and reported to the host computer. The host computer restarts, repairs, replaces and other operations according to the equipment fault situation. When the laser radar self-test is normal, obstacle detection and positioning are performed.

[0065] In the specific implementation scenario, the laser radar adopts a composite sensing technology route, integrating detection perception, loose deflection, dust, and occlusion. It has a built-in MEMS attitude sensor. When it detects that the device is tilted forward and backward, left and right at an angle of more than 3 degrees, and the up and down offset is 20mm, the system prompts the device-related alarm. The laser radar has a built-in temperature sensor. When the temperature of the radar core components or circuit elements is detected to be abnormal, the system maintenance prompt is given in time. The laser radar also supports self-diagnosis dust detection. When the lens is contaminated or the accumulation of dirt exceeds the positioning tolerance, the detection system can issue a maintenance alarm signal, and will not output an obstacle detection false alarm.

[0066] In some embodiments of the present application, dust occlusion detection specifically includes:

[0067] Determine a data point within a first preset distance from the data origin as a dust point, and issue a dust prompt if the dust point exists;

[0068] Determine a data point between the first preset distance and the second preset distance from the data origin as an occlusion point, and issue an occlusion prompt if the occlusion point exists;

[0069] Wherein, the second preset distance is greater than the first preset distance.

[0070] If the laser radar is in an abnormal external working environment such as dust, obstruction, etc., the host computer program will be notified in time to prompt manual maintenance and fault handling. Obstacle detection and positioning will only be performed when the laser radar is in a normal working environment. Usually the first preset distance is 10cm and the second preset distance is 50cm, which can also be set according to the actual application scenario.

[0071] In this embodiment, determining whether the current detection boundary intrudes into the defense zone boundary specifically includes:

[0072] The data coordinates of the current detection boundary are located within the defense zone boundary and the data point outside the second preset distance from the data origin is determined as an obstacle point. If the obstacle point exists, it means that the current detection boundary invades the defense zone boundary.

[0073] In this embodiment, if there is no obstacle in the defense zone, the detection and analysis will continue until the detection countdown ends; if there is an obstacle in the defense zone, the obstacle alarm information will be uploaded to the host computer. The detection result can be intuitively viewed through the mobile terminal APP to see the obstacle intrusion and intrusion position. At the same time, the edge computing module of the laser radar can perform cluster analysis on the scanned object based on the scan of the detected object, the shape, size, and clamped position of the obstacle, so as to accurately determine the type of the clamped object and the risk level of the clamped foreign object, and effectively filter out foreign objects such as mosquitoes and ground garbage to ensure the vehicle's operating efficiency.

[0074] By adopting the above technical solution disclosed in the present invention, the following beneficial effects are obtained:

[0075] The present invention adopts a ToF-based laser radar with an IP67 protection level and no installation restrictions; the regional laser light curtain sector scanning detection can achieve full coverage of the gap between the platform door and the car door; the detection accuracy is 10-20mm, and precise positioning can be achieved. The laser radar front-end embedded edge computing module can independently complete the defense zone learning and background interference removal functions without relying on the back-end controller through the front-end stabilization algorithm, and can directly judge the situation of people and objects being clamped in the gap between the platform door and the car door. The laser radar is integrated with transceiver, resistant to vibration interference, and easy to debug; it adopts a 905nm wavelength beam to resist strong light interference; it supports dust and occlusion detection alarms; it supports linkage with signals, platform doors, and video systems; dual-channel CAN bus, supports 100% concurrent alarms for multiple doors, supports single-door fault isolation, and emergency bypass of the entire side platform. The method provided by the present invention has strong environmental applicability (applicable to indoor, outdoor, straight, curved, half-height, full-height platforms and fully automatic driving lines).

[0076] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be considered as the scope of protection of the present invention.

Claims

1. A method for detecting obstacles in the gap of a shielding door based on edge computing laser radar, characterized in that: Applied to a ToF-based laser radar, the laser radar is provided with an edge computing module, and the method includes: Scanning the environment through the laser radar to collect current detection data; Processing the current detection data by the edge computing module to obtain a current detection boundary; Determine whether the current detection boundary intrudes into the defense zone boundary; if so, it indicates that there is an obstacle in the defense zone, and the obstacle alarm information is uploaded; if not, it indicates that there is no obstacle in the defense zone; Wherein, the data processing includes through filtering processing, conditional filtering processing, data conversion processing and statistical filtering processing in sequence, so as to filter background interference data; The data processing specifically includes: By presetting the detection angle, the data exceeding the preset detection angle is filtered; By presetting the detection radius, the data exceeding the preset detection radius is filtered; Convert the filtered valid data into two-dimensional polar coordinates of detection distance and detection angle and store them; Based on the two-dimensional polar coordinates, the average detection distance of each data point in multiple detection cycles and the change amount relative to adjacent data points are statistically analyzed, and the data points whose average detection distance exceeds the statistical mean and variance preset thresholds and the data points whose change amount exceeds the standard deviation preset threshold are filtered.

2. The method according to claim 1, characterized in that The defense zone boundary is obtained by the following method: Scanning an obstacle-free environment by using the laser radar to collect original detection data; Performing the data processing on the original detection data by the edge computing module to obtain a detection boundary; Based on the data coordinates of the detection boundary, the defense zone boundary is obtained by shrinking the preset safety distance toward the data origin.

3. The method according to claim 1, characterized in that The laser radar is also provided with a wireless communication device, which is connected to a mobile terminal through the wireless communication device, and the detection results are displayed and the parameters of the laser radar are set through the mobile terminal.

4. The method according to claim 3, characterized in that The laser radar is also provided with a temperature sensor and a posture sensor; before the laser radar scans, it also performs a status self-check, and when any inspection result does not meet the preset standard, a maintenance prompt is issued; the inspection items of the status self-check include: the temperature measured by the temperature sensor, the three-dimensional posture information measured by the posture sensor, the communication status and the storage status.

5. The method according to claim 4, characterized in that The preset standard of the three-dimensional posture information is the three-dimensional posture information when the laser radar obtains the defense zone boundary, and the three-dimensional posture information includes pitch angle, left and right side angle and vertical rotation angle.

6. The method according to claim 1, characterized in that Before determining whether the current detection boundary intrudes into the defense zone boundary, a dust cover detection is also performed, including: Determine a data point within a first preset distance from the data origin as a dust point, and issue a dust prompt if the dust point exists; Determine a data point between the first preset distance and the second preset distance from the data origin as an occlusion point, and issue an occlusion prompt if the occlusion point exists; Wherein, the second preset distance is greater than the first preset distance.

7. The method according to claim 6, characterized in that Determining whether the current detection boundary intrudes into the defense zone boundary specifically includes: The data coordinates of the current detection boundary are located within the defense zone boundary and the data point outside the second preset distance from the data origin is determined as an obstacle point. If the obstacle point exists, it means that the current detection boundary invades the defense zone boundary.

8. The method according to claim 7, characterized in that If the current detection boundary intrudes into the defense zone boundary, it also includes obtaining the shape, size and position of the corresponding obstacle based on the data information of the obstacle point, and performing clustering restoration to identify the corresponding risk level and generate obstacle alarm information.

Images