Train detection method, device, electronic equipment and storage medium

Abstract

The present application relates to the field of rail transit technology, and in particular to a train detection method, device, electronic device and storage medium. The present application can identify the train point cloud data of the second train from the scene point cloud data by acquiring the scene image in front of the first train. Then, based on the train point cloud data, the train profile of the second train can be determined, and based on the train profile, the relative position and relative travel direction between the first train and the second train can be calculated. In this way, by combining the scene image in front of the first train and the scene point cloud data, the relative position and relative travel direction between the first train and the second train can be accurately calculated. The operation is simple and can improve the safety of train travel.

Description

Technical Field

[0001] The present application relates to the field of rail transit technology, and in particular to a train detection method, device, electronic equipment and storage medium. Background Art

[0002] Typically, different trains travel on different tracks. However, when they reach an intersection, the tracks intersect and the current train may have a direct plane conflict with trains on other tracks. Therefore, it is necessary to detect and identify the train in front to determine whether there is a possibility of collision with the current train. Specifically, the relative position and relative travel direction between the current train and the train in front are calculated to determine whether there is a possibility of collision between the current train and the train in front.

[0003] Traditional rail transit technology generally detects trains by configuring train sensing equipment on the ground. The installation of this equipment is relatively complex and cannot guarantee the accuracy of detecting the train ahead. Summary of the Invention

[0004] In view of this, the embodiments of the present application provide at least one train detection method, device, electronic device and storage medium, which are easy to operate and can improve the safety of train travel.

[0005] This application mainly includes the following aspects:

[0006] In a first aspect, an embodiment of the present application provides a train detection method, the train detection method comprising:

[0007] Acquire scene images and scene point cloud data in front of the first train;

[0008] Based on the scene image, identifying train point cloud data of a second train from the scene point cloud data; wherein the second train is located in front of the first train;

[0009] determining a train profile of the second train based on the train point cloud data;

[0010] Based on the train profiles, a relative position and a relative travel direction between the first train and the second train are calculated.

[0011] In a possible implementation, identifying train point cloud data of a second train from the scene point cloud data based on the scene image includes:

[0012] Identifying a region image corresponding to the second train from the scene image, and determining, based on the region image, region point cloud data containing the second train from the scene point cloud data;

[0013] Train point cloud data of the second train is identified from the regional point cloud data.

[0014] In a possible implementation, determining, based on the regional image, regional point cloud data containing the second train from the scene point cloud data includes:

[0015] The scene point cloud data is projected onto the scene image, and the region point cloud data projected onto the region image is acquired from the scene point cloud data.

[0016] In a possible implementation, train point cloud data of the second train is identified from the regional point cloud data according to the following steps:

[0017] Clustering the regional point cloud data according to the train feature information to obtain a plurality of clustered point cloud data;

[0018] The train point cloud data of the second train is filtered out from the plurality of aggregated point cloud data.

[0019] In a possible implementation, before clustering the regional point cloud data according to the train characteristic information to obtain a plurality of clustered point cloud data, the train detection method further includes:

[0020] The regional point cloud data is rasterized to obtain rasterized regional point cloud data.

[0021] In a possible implementation, filtering out the train point cloud data of the second train from the plurality of aggregated point cloud data includes:

[0022] Determining, from the plurality of aggregated point cloud data, point cloud data of a first region corresponding to each train segment and point cloud data of a second region corresponding to each non-train segment;

[0023] The train point cloud data of the second train is filtered out from the plurality of aggregated point cloud data according to a positional relationship between the first area and the second area.

[0024] In one possible implementation, the relative position between the first train and the second train is calculated according to the following steps:

[0025] Extracting the train head contour points from the train contour;

[0026] calculating a target distance between the second train and the first train based on the locomotive contour points;

[0027] The relative position between the first train and the second train is determined according to the target distance.

[0028] In one possible implementation, the relative travel direction between the first train and the second train is calculated according to the following steps:

[0029] Extracting a train head contour point and at least two carriage contour points from the train contour;

[0030] Fitting the vehicle head contour points and the at least two vehicle body contour points into a target curve;

[0031] The relative travel direction between the first train and the second train is determined according to the target curve.

[0032] In a second aspect, an embodiment of the present application further provides a train detection device, the train detection device comprising:

[0033] an acquisition module, configured to acquire scene images and scene point cloud data in front of the first train;

[0034] an identification module, configured to identify train point cloud data of a second train from the scene point cloud data based on the scene image; wherein the second train is located in front of the first train;

[0035] a determination module, configured to determine a train profile of the second train based on the train point cloud data;

[0036] A calculation module is used to calculate the relative position and relative travel direction between the first train and the second train based on the train profile.

[0037] In a possible implementation, the identification module includes:

[0038] a determining unit, configured to identify a region image corresponding to the second train from the scene image, and determine, based on the region image, region point cloud data containing the second train from the scene point cloud data;

[0039] An identification unit is used to identify the train point cloud data of the second train from the regional point cloud data.

[0040] In a possible implementation, the determining unit is configured to determine the regional point cloud data from the scene point cloud data according to the following conditions:

[0041] The scene point cloud data is projected onto the scene image, and the region point cloud data projected onto the region image is acquired from the scene point cloud data.

[0042] In a possible implementation, the identification unit is configured to identify the train point cloud data according to the following steps:

[0043] Clustering the regional point cloud data according to the train feature information to obtain a plurality of clustered point cloud data;

[0044] The train point cloud data of the second train is filtered out from the plurality of aggregated point cloud data.

[0045] In a possible implementation, the identification module further includes:

[0046] The processing unit is used to perform rasterization processing on the regional point cloud data to obtain rasterized regional point cloud data.

[0047] In a possible implementation, the processing unit is configured to filter out the train point cloud data according to the following requirements:

[0048] Determining, from the plurality of aggregated point cloud data, point cloud data of a first region corresponding to each train segment and point cloud data of a second region corresponding to each non-train segment;

[0049] The train point cloud data of the second train is filtered out from the plurality of aggregated point cloud data according to a positional relationship between the first area and the second area.

[0050] In a possible implementation, the calculation module is configured to calculate the relative position between the first train and the second train according to the following steps:

[0051] Extracting the train head contour points from the train contour;

[0052] calculating a target distance between the second train and the first train based on the locomotive contour points;

[0053] The relative position between the first train and the second train is determined according to the target distance.

[0054] In a possible implementation, the calculation module is configured to calculate the relative travel direction between the first train and the second train according to the following steps:

[0055] Extracting a train head contour point and at least two carriage contour points from the train contour;

[0056] Fitting the vehicle head contour points and the at least two vehicle body contour points into a target curve;

[0057] The relative travel direction between the first train and the second train is determined according to the target curve.

[0058] In a third aspect, an embodiment of the present application further provides an electronic device comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the train detection method described in the first aspect or any possible implementation method of the first aspect.

[0059] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the train detection method described in the first aspect or any possible implementation method of the first aspect are executed.

[0060] The train detection method, device, electronic device and storage medium provided in the embodiments of the present application use the acquired scene image in front of the first train to identify the train point cloud data of the second train from the scene point cloud data. Based on the train point cloud data, the train profile of the second train can be determined, and based on the train profile, the relative position and relative travel direction between the first train and the second train can be calculated. Compared with the existing technology of detecting trains by generally configuring train sensing equipment on the ground, which is relatively complex to lay out and cannot guarantee the accuracy of detecting the train in front, this method is simple to operate and can improve the safety of train travel.

[0061] Furthermore, the train detection method provided in the embodiments of the present application determines, from multiple aggregated point cloud data, point cloud data for a first region corresponding to each train segment and point cloud data for a second region corresponding to each non-train segment. Based on the positional relationship between the first and second regions, the train point cloud data for the second train is filtered out from the multiple aggregated point cloud data. This allows accurate filtering of the train point cloud data for the second train, thereby improving train safety.

[0062] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0064] Figure 1 A flow chart of a train detection method provided in an embodiment of the present application is shown;

[0065] Figure 2 A functional module diagram of a train detection device provided in an embodiment of the present application is shown;

[0066] Figure 3 A schematic diagram of the structure of the identification module provided in an embodiment of the present application is shown;

[0067] Figure 4 A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown.

[0068] Description of main component symbols:

[0069] In the figure: 200 - train detection device; 210 - acquisition module; 220 - identification module; 221 - determination unit; 222 - identification unit; 223 - processing unit; 230 - determination module; 240 - calculation module; 400 - electronic device; 410 - processor; 420 - memory; 430 - bus. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of illustration and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowcharts can be implemented out of sequence, and steps without logical context can be reversed or implemented simultaneously. In addition, those skilled in the art, under the guidance of the contents of this application, can add one or more other operations to the flowchart, or remove one or more operations from the flowchart.

[0071] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.

[0072] In order to enable those skilled in the art to use the contents of this application, the following implementation methods are given in combination with the specific application scenario "train collision avoidance". For those skilled in the art, the general principles defined here can be applied to other embodiments and application scenarios without departing from the spirit and scope of this application.

[0073] The following methods, devices, electronic devices or computer-readable storage media of the embodiments of the present application can be applied to any scenario requiring train detection. The embodiments of the present application are not limited to specific application scenarios. Any scheme using the train detection method and device provided by the embodiments of the present application is within the scope of protection of this application.

[0074] It is worth noting that before this application was submitted, traditional rail transit technology generally detected trains by configuring train sensing equipment on the ground. The installation of this equipment is relatively complex and cannot guarantee the accuracy of detecting the train ahead.

[0075] To address the above issues, the embodiments of the present application can identify the train point cloud data of the second train from the scene point cloud data by acquiring the scene image in front of the first train. Furthermore, based on the train point cloud data, the train profile of the second train can be determined, and based on the train profile, the relative position and relative travel direction between the first and second trains can be calculated. In this way, by combining the scene image in front of the first train with the scene point cloud data, the relative position and relative travel direction between the first and second trains can be accurately calculated, which is simple to operate and can improve the safety of train travel.

[0076] To facilitate understanding of the present application, the technical solutions provided in the present application are described in detail below in conjunction with specific embodiments.

[0077] Figure 1 This is a flow chart of a train detection method provided in an embodiment of the present application. Figure 1 As shown, the train detection method provided in the embodiment of the present application includes the following steps:

[0078] S101: Acquire scene images and scene point cloud data in front of a first train.

[0079] In a specific implementation, the scene image and scene point cloud data in front of the first train (current train) can be obtained in real time, or the scene image and scene point cloud data in front of the current train can be obtained at preset time intervals, and the scene image and scene point cloud data in front of the current train can be obtained when a preset distance from an intersection is detected. Here, the preset time duration can be set according to the emission frequency of the lidar, and the preset distance can be set according to the average speed of the train.

[0080] It should be noted that the scene image in front of the first train can be collected by a camera, and the scene point cloud data in front of the first train can be collected by a lidar, wherein the camera and the lidar are both installed on the front of the first train to collect the scene data in front of the first train.

[0081] Here, LiDAR is a measurement device that combines laser scanning with a positioning and attitude determination system. The LiDAR system includes a laser and a receiving system. The laser generates and emits a beam of light pulses, which hits the object and reflects back, and is eventually received by the receiver. The receiver accurately measures the propagation time of the light pulse from emission to reflection. The propagation time can be converted into a distance measurement. Combined with the height of the laser and the laser scanning angle, the three-dimensional coordinates X, Y, and Z of each ground light spot can be accurately calculated. Point cloud data refers to a set of vectors in a three-dimensional coordinate system. These vectors are usually expressed in the form of X, Y, and Z three-dimensional coordinates, and are generally used to represent the outer surface shape of an object. LiDAR point cloud data is obtained by LiDAR scanning. While the point cloud can be directly depicted and observed, the point cloud itself cannot usually be used directly for 3D applications. Therefore, it is generally converted into a mesh model such as polygons or triangles through surface reconstruction methods.

[0082] S102: Based on the scene image, identify train point cloud data of a second train from the scene point cloud data; wherein the second train is located in front of the first train.

[0083] In a specific implementation, after obtaining the scene image and scene point cloud data in front of the first train, the train point cloud data of the second train located in front of the first train can be identified from the scene point cloud data based on the scene image. Here, the scene image includes the image of the second train, and the scene point cloud data includes the train point cloud data of the second train.

[0084] Furthermore, a regional image containing the second train may be firstly captured from the scene image, and then regional point cloud data of the second train corresponding to the regional image may be determined from the scene point cloud data. Furthermore, train point cloud data of the second train may be identified from the regional point cloud data. Specifically, step S102 includes the following steps of identifying the train point cloud data of the second train from the scene point cloud data based on the scene image:

[0085] Step a: Identify the area image corresponding to the second train from the scene image, and based on the area image, determine the area point cloud data containing the second train from the scene point cloud data.

[0086] In a specific implementation, an object detection algorithm can be used to identify the regional image corresponding to the second train from the scene image. Since the scene image and the scene point cloud data correspond to each other, the regional point cloud data corresponding to the regional image and containing the second train can be determined from the scene point cloud data. Here, the regional image is a portion of the scene image that contains the image of the second train. Similarly, the regional point cloud data is a portion of the scene point cloud data that contains the point cloud data of the second train.

[0087] Here, object detection, based on image classification, can be implemented using deep learning or other computer vision techniques. Object detection can accurately locate the position of each object in an image. Currently, mainstream object detection algorithms fall into two main categories: one is the evolution of object detection algorithms (two-stage detection algorithms), such as the Region-Convolutional Neural Network (RCNN) family of algorithms. The main idea is to first generate a series of sparse candidate bounding boxes using heuristic methods or CNN networks, and then perform classification and regression on these candidate boxes. The advantage of two-stage methods is high accuracy. The other is one-stage methods, such as YOLO and Single Shot Detection (SSD). The main idea is to uniformly and densely sample different locations in the image, using different scales and aspect ratios. The features are then extracted using a CNN, followed by direct classification and regression. The entire process only requires one step, resulting in high speed. However, the disadvantage of uniform and dense sampling is that it is difficult to train, mainly due to the imbalance between positive and negative samples (background), resulting in low model accuracy.

[0088] Among them, the YOLO algorithm only uses a CNN to directly predict the category and location of different targets. The YOLO algorithm uses a separate CNN model to achieve end-to-end target detection. First, the input image is sent to the CNN network, and finally the network prediction results are processed to obtain the detected target.

[0089] Optionally, the process of determining the regional point cloud data containing the second train from the scene point cloud data is described below. That is, in step a, determining the regional point cloud data containing the second train from the scene point cloud data based on the regional image includes the following steps:

[0090] The scene point cloud data is projected onto the scene image, and the region point cloud data projected onto the region image is acquired from the scene point cloud data.

[0091] In a specific implementation, the 3D scene point cloud data can be first projected onto a 2D scene image. Here, the scene image marks the image region where the second train is located, that is, the regional image corresponding to the second train is marked. Then, the regional point cloud data projected onto the regional image can be directly obtained from the scene point cloud data. Here, the regional point cloud data is 3D point cloud data. In this way, by projecting the 3D point cloud data onto the 2D image plane based on the joint calibration results of the camera and lidar, the point cloud data corresponding to the image pixel region can be extracted.

[0092] Step b: Identify the train point cloud data of the second train from the regional point cloud data.

[0093] In a specific implementation, after determining the regional point cloud data containing the second train from the scene point cloud data, further, the train point cloud data of the second train is identified from the regional point cloud data, that is, the train point cloud data of the second train is the point cloud data from the regional point cloud data excluding the point cloud data that does not belong to the second train.

[0094] Furthermore, the process of identifying the train point cloud data of the second train from the regional point cloud data in step b is described below:

[0095] Step b11: performing rasterization processing on the regional point cloud data to obtain rasterized regional point cloud data.

[0096] In the specific implementation, the regional point cloud data is first rasterized to obtain the rasterized regional point cloud data. In this way, the use of the rasterized regional point cloud data can reduce the complexity of the calculation, and the train point cloud data of the second train can be identified more quickly from the rasterized regional point cloud data.

[0097] Here, the rasterization process discards the height information of the three-dimensional point cloud data, wherein the height information is stored separately, and the three-dimensional point cloud data is projected onto a plane at the same height. The process is to reduce from three dimensions to two dimensions. In this way, the amount of calculation for calculating the train point cloud data of the second train can be reduced, thereby improving the calculation efficiency.

[0098] Step b12: clustering the rasterized regional point cloud data according to the train feature information to obtain a plurality of clustered point cloud data.

[0099] In a specific implementation, the point cloud data in the regional point cloud data is clustered according to the train feature information to obtain a plurality of clustered point cloud data, wherein each clustered point cloud data is formed by clustering a plurality of point cloud data together, and the type of the point cloud data in each clustered point cloud data is the same, that is, each clustered point cloud data can be train point cloud data or non-train point cloud data.

[0100] Here, the train feature information includes the length feature, width feature, etc. of the train. Through the train feature information, the point cloud data of the obstruction in the regional point cloud data and the point cloud data of the second train can be identified.

[0101] Step b13: Filtering out the train point cloud data of the second train from the plurality of aggregated point cloud data.

[0102] In a specific implementation, the train point cloud data of the second train can be screened out from the multiple aggregated point cloud data.

[0103] Optionally, the process of screening out the train point cloud data of the second train from the plurality of aggregated point cloud data in step b13 is described below:

[0104] From the multiple aggregated point cloud data, determine the point cloud data of the first area corresponding to each train segment and the point cloud data of the second area corresponding to each non-train segment; based on the positional relationship between the first area and the second area, filter out the train point cloud data of the second train from the multiple aggregated point cloud data.

[0105] In a specific implementation, after obtaining a plurality of aggregated point cloud data, each aggregated point cloud data is composed of a plurality of point cloud data of the same type. Since the second train may be blocked by one or more obstructions, the point cloud data of the first area corresponding to each train section and the point cloud data of the second area corresponding to each non-train (obstruction) can be determined from the plurality of aggregated point cloud data. In this way, the train point cloud data belonging to the second train can be determined from the plurality of aggregated point cloud data according to the characteristics of the train, and the train point cloud data of the second train can be marked in the regional point cloud data according to the position of the first area corresponding to each train section in the regional point cloud data, and the positional relationship between each first area and the second area. Here, according to the characteristics of the train, by detecting the obstructions blocking the second train, the problem of inaccurate train point cloud data detection caused by the obstruction of the second train can be solved, thereby improving the accuracy of train point cloud data detection.

[0106] It should be noted that the point cloud data contained in the preceding train (the second train) is detected based on the train characteristics, including the detection of obstacles between the lidar and the preceding train that would block the lidar from scanning the preceding train. If there are n obstacles that meet the conditions, the preceding train will be divided into n+1 segments, thereby detecting all the point cloud data of the train and detecting the preceding train.

[0107] S103: Determine a train profile of the second train based on the train point cloud data.

[0108] In the specific implementation, contour points are extracted from the detected train point cloud data, and then the train contour of the second train is obtained. Specifically, the detected train point cloud data is downsampled to reduce the complexity of the algorithm, and the contour point detection algorithm is used to extract contour points from the downsampled train point cloud data to obtain the contour of the front train.

[0109] S104: Calculate the relative position and relative travel direction between the first train and the second train based on the train profile.

[0110] In a specific implementation, after obtaining the train profile of the second train, the relative position and relative travel direction between the first train (the current train) and the second train can be calculated based on the train profile. In this way, it can be accurately determined whether the first train and the second train will collide. Then, based on the judgment result, the travel speed of the first train can be controlled to improve the safety of the train travel.

[0111] Optionally, the relative position between the first train and the second train is calculated according to the following steps:

[0112] Extracting locomotive contour points from the train contour; calculating a target distance between the second train and the first train based on the locomotive contour points; and determining a relative position between the first train and the second train based on the target distance.

[0113] In practice, the train's head contour points are first extracted from the train's outline. Typically, a laser radar (LIDAR) is installed on the first train, and the LIDAR serves as the coordinate origin. Therefore, the target distance between the second train and the first train can be directly calculated based on the coordinates of the head contour points. Furthermore, the relative position between the first and second trains is determined based on the target distance. Here, the relative position refers to the relative distance between the first and second trains.

[0114] Optionally, the relative position between the first train and the second train is calculated according to the following steps:

[0115] Extracting a locomotive contour point and at least two carriage contour points from the train contour; fitting the locomotive contour point and the at least two carriage contour points into a target curve; and determining the relative travel direction between the first train and the second train based on the target curve.

[0116] In a specific implementation, the locomotive contour point and at least two carriage contour points are extracted from the train contour, and the extracted locomotive contour point and the at least two contour points are fitted into a target curve. According to the curve direction of the curve, the relative driving direction between the first train and the second train is determined. Here, the relative driving direction includes a relatively opposite direction and a relatively same direction, wherein the relatively opposite direction is not an absolutely opposite direction, and the relatively same direction is not an absolutely same direction.

[0117] It should be noted that traditional rail transit technology generally detects trains by configuring train sensing equipment on the ground. The installation of this equipment is relatively complex and cannot guarantee the accuracy of detecting the train in front. In this regard, the present application combines the scene image in front of the first train and the scene point cloud data to accurately calculate the relative position and relative travel direction between the first train and the second train. It is easy to operate and can improve the safety of train travel.

[0118] In an embodiment of the present application, by acquiring a scene image in front of the first train, the train point cloud data of the second train can be identified from the scene point cloud data. Furthermore, based on the train point cloud data, the train profile of the second train can be determined, and based on the train profile, the relative position and relative travel direction between the first and second trains can be calculated. In this way, by combining the scene image in front of the first train with the scene point cloud data, the relative position and relative travel direction between the first and second trains can be accurately calculated, which is simple to operate and can improve the safety of train travel.

[0119] Based on the same application concept, the embodiments of the present application also provide a train detection device corresponding to the train detection method provided in the above embodiments. Since the principle of solving the problem by the device in the embodiments of the present application is similar to the train detection method in the above embodiments of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0120] Figure 2 A functional module diagram of a train detection device 200 provided in an embodiment of the present application is shown; Figure 3 A schematic structural diagram of the identification module 220 provided in an embodiment of the present application is shown.

[0121] like Figure 2 As shown, the train detection device 200 includes:

[0122] An acquisition module 210 is configured to acquire a scene image and scene point cloud data in front of the first train;

[0123] an identification module 220 for identifying train point cloud data of a second train from the scene point cloud data based on the scene image; wherein the second train is located in front of the first train;

[0124] a determination module 230, configured to determine a train profile of the second train based on the train point cloud data;

[0125] The calculation module 240 is configured to calculate the relative position and relative travel direction between the first train and the second train based on the train profile.

[0126] In one possible implementation, Figure 3 As shown, the identification module 220 includes:

[0127] a determining unit 221 configured to identify a region image corresponding to the second train from the scene image, and determine, based on the region image, region point cloud data containing the second train from the scene point cloud data;

[0128] The identification unit 222 is configured to identify the train point cloud data of the second train from the regional point cloud data.

[0129] In one possible implementation, Figure 3 As shown, the determining unit 221 is configured to determine the regional point cloud data from the scene point cloud data according to the following conditions:

[0130] The scene point cloud data is projected onto the scene image, and the region point cloud data projected onto the region image is acquired from the scene point cloud data.

[0131] In one possible implementation, Figure 3 As shown, the identification unit 222 is used to identify the train point cloud data according to the following steps:

[0132] Clustering the regional point cloud data according to the train feature information to obtain a plurality of clustered point cloud data;

[0133] The train point cloud data of the second train is filtered out from the plurality of aggregated point cloud data.

[0134] In one possible implementation, Figure 3 As shown, the identification module 220 further includes:

[0135] The processing unit 223 is configured to perform rasterization processing on the regional point cloud data to obtain rasterized regional point cloud data.

[0136] In one possible implementation, Figure 3 As shown, the processing unit 223 is used to filter out the train point cloud data according to the following requirements:

[0137] Determining, from the plurality of aggregated point cloud data, point cloud data of a first region corresponding to each train segment and point cloud data of a second region corresponding to each non-train segment;

[0138] The train point cloud data of the second train is filtered out from the plurality of aggregated point cloud data according to a positional relationship between the first area and the second area.

[0139] In one possible implementation, Figure 2 As shown, the calculation module 240 is used to calculate the relative position between the first train and the second train according to the following steps:

[0140] Extracting the train head contour points from the train contour;

[0141] calculating a target distance between the second train and the first train based on the locomotive contour points;

[0142] The relative position between the first train and the second train is determined according to the target distance.

[0143] In one possible implementation, Figure 2 As shown, the calculation module 240 is used to calculate the relative travel direction between the first train and the second train according to the following steps:

[0144] Extracting a train head contour point and at least two carriage contour points from the train contour;

[0145] Fitting the vehicle head contour points and the at least two vehicle body contour points into a target curve;

[0146] The relative travel direction between the first train and the second train is determined according to the target curve.

[0147] In the embodiments of the present application, by acquiring a scene image in front of the first train, the train point cloud data of the second train can be identified from the scene point cloud data. Furthermore, based on the train point cloud data, the train profile of the second train can be determined, and based on the train profile, the relative position and relative travel direction between the first and second trains can be calculated. In this way, by combining the scene image in front of the first train with the scene point cloud data, the relative position and relative travel direction between the first and second trains can be accurately calculated, which is simple to operate and can improve the safety of train travel.

[0148] Based on the same application concept, see Figure 4As shown, it is a structural diagram of an electronic device 400 provided in an embodiment of the present application, including: a processor 410, a memory 420 and a bus 430, wherein the memory 420 stores machine-readable instructions executable by the processor 410, and when the electronic device 400 is running, the processor 410 and the memory 420 communicate with each other through the bus 430, and the machine-readable instructions are executed by the processor 410 when running to perform the steps of the train detection method as described in any of the above embodiments.

[0149] Specifically, when the machine-readable instructions are executed by the processor 410, the following processing may be performed:

[0150] Acquire scene images and scene point cloud data in front of the first train;

[0151] Based on the scene image, identifying train point cloud data of a second train from the scene point cloud data; wherein the second train is located in front of the first train;

[0152] determining a train profile of the second train based on the train point cloud data;

[0153] Based on the train profiles, a relative position and a relative travel direction between the first train and the second train are calculated.

[0154] Optionally, when the machine-readable instructions are executed by the processor 410, the following processing may be performed:

[0155] Identifying a region image corresponding to the second train from the scene image, and determining, based on the region image, region point cloud data containing the second train from the scene point cloud data;

[0156] Train point cloud data of the second train is identified from the regional point cloud data.

[0157] Optionally, when the machine-readable instructions are executed by the processor 410, the following processing may be performed:

[0158] The scene point cloud data is projected onto the scene image, and the region point cloud data projected onto the region image is acquired from the scene point cloud data.

[0159] Optionally, when the machine-readable instructions are executed by the processor 410, the following processing may be performed:

[0160] Clustering the regional point cloud data according to the train feature information to obtain a plurality of clustered point cloud data;

[0161] The train point cloud data of the second train is filtered out from the plurality of aggregated point cloud data.

[0162] Optionally, when the machine-readable instructions are executed by the processor 410, the following processing may be performed:

[0163] The regional point cloud data is rasterized to obtain rasterized regional point cloud data.

[0164] Optionally, when the machine-readable instructions are executed by the processor 410, the following processing may be performed:

[0165] Determining, from the plurality of aggregated point cloud data, point cloud data of a first region corresponding to each train segment and point cloud data of a second region corresponding to each non-train segment;

[0166] The train point cloud data of the second train is filtered out from the plurality of aggregated point cloud data according to a positional relationship between the first area and the second area.

[0167] Optionally, when the machine-readable instructions are executed by the processor 410, the following processing may be performed:

[0168] Extracting the train head contour points from the train contour;

[0169] calculating a target distance between the second train and the first train based on the locomotive contour points;

[0170] The relative position between the first train and the second train is determined according to the target distance.

[0171] Optionally, when the machine-readable instructions are executed by the processor 410, the following processing may be performed:

[0172] Extracting a train head contour point and at least two carriage contour points from the train contour;

[0173] Fitting the vehicle head contour points and the at least two vehicle body contour points into a target curve;

[0174] The relative travel direction between the first train and the second train is determined according to the target curve.

[0175] In an embodiment of the present application, by acquiring a scene image in front of the first train, the train point cloud data of the second train can be identified from the scene point cloud data. Furthermore, based on the train point cloud data, the train profile of the second train can be determined, and based on the train profile, the relative position and relative travel direction between the first and second trains can be calculated. In this way, by combining the scene image in front of the first train with the scene point cloud data, the relative position and relative travel direction between the first and second trains can be accurately calculated, which is simple to operate and can improve the safety of train travel.

[0176] Based on the same application concept, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the train detection method provided in the above embodiment are executed.

[0177] Specifically, the storage medium can be a general storage medium, such as a mobile disk, a hard disk, etc. When the computer program on the storage medium is run, the above-mentioned train detection method can be executed. By combining the scene image in front of the first train and the scene point cloud data, the relative position and relative travel direction between the first train and the second train can be accurately calculated. The operation is simple and can improve the safety of train travel.

[0178] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0179] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0180] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0181] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0182] The above are only specific embodiments of the present application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A train detection method, characterized in that: The train detection method comprises: Acquire scene images and scene point cloud data in front of the first train; Based on the scene image, identifying train point cloud data of a second train from the scene point cloud data; wherein the second train is located in front of the first train; determining a train profile of the second train based on the train point cloud data; calculating a relative position and a relative travel direction between the first train and the second train based on the train profile; Among them, the identifying of the train point cloud data of the second train from the scene point cloud data based on the scene image includes: identifying the area image corresponding to the second train from the scene image, projecting the scene point cloud data onto the scene image, and obtaining the area point cloud data projected on the area image from the scene point cloud data; rasterizing the area point cloud data, discarding the height information, and obtaining rasterized area point cloud data; clustering the rasterized area point cloud data according to train feature information to obtain a plurality of aggregated point cloud data; determining the point cloud data of the first area corresponding to each train and the point cloud data of the second area corresponding to each non-train from the plurality of aggregated point cloud data; and filtering out the train point cloud data of the second train from the plurality of aggregated point cloud data according to the positional relationship between the first area and the second area.

2. The train detection method according to claim 1, characterized in that: The relative position between the first train and the second train is calculated according to the following steps: Extracting the train head contour points from the train contour; calculating a target distance between the second train and the first train based on the locomotive contour points; The relative position between the first train and the second train is determined according to the target distance.

3. The train detection method according to claim 1, characterized in that: The relative travel direction between the first train and the second train is calculated according to the following steps: Extracting a train head contour point and at least two carriage contour points from the train contour; Fitting the vehicle head contour points and the at least two vehicle body contour points into a target curve; The relative travel direction between the first train and the second train is determined according to the target curve.

4. A train detection device, characterized in that: The train detection device comprises: an acquisition module, configured to acquire scene images and scene point cloud data in front of the first train; an identification module, configured to identify train point cloud data of a second train from the scene point cloud data based on the scene image; wherein the second train is located in front of the first train; a determination module, configured to determine a train profile of the second train based on the train point cloud data; a calculation module, configured to calculate a relative position and a relative travel direction between the first train and the second train based on the train profile; Among them, the recognition module includes a determination unit, an identification unit and a processing unit; the determination unit is used to identify the area image corresponding to the second train from the scene image, and project the scene point cloud data onto the scene image, and obtain the area point cloud data projected on the area image from the scene point cloud data; the processing unit is used to rasterize the area point cloud data, discard the height information, and obtain rasterized area point cloud data; the recognition unit is used to cluster the rasterized area point cloud data according to train feature information to obtain multiple aggregated point cloud data; the processing unit is used to determine the point cloud data of the first area corresponding to each train and the point cloud data of the second area corresponding to each non-train from the multiple aggregated point cloud data; according to the positional relationship between the first area and the second area, the train point cloud data of the second train is filtered out from the multiple aggregated point cloud data.

5. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the train detection method as described in any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the train detection method according to any one of claims 1 to 3 are executed.

Images