A method, device, equipment and storage medium for detecting foreign objects in railway track area

Abstract

The present invention discloses a method, device, equipment and storage medium for detecting foreign objects in a railway track area, wherein the method includes the steps of: performing point cloud segmentation on point cloud data of a railway environment to obtain a foreign object label set and a track label set; extracting point cloud data of the track area according to the track label set; comparing the point cloud data of the track area of ​​the previous and next frames to obtain a newly added point cloud data set; performing cluster analysis on the newly added point cloud data set to obtain a foreign object cluster; cross-calculating the foreign object cluster with the foreign object label set to calculate and identify foreign objects in the track area. The method of the present application accurately identifies and extracts point cloud data of the track area through point cloud segmentation, and identifies foreign objects in the track area by cross-calculating the foreign object cluster with the foreign object label set, which can not only reduce the interference of environmental changes on foreign object detection, but also improve detection accuracy and detection efficiency.

Description

Technical Field

[0001] The present invention relates to the field of foreign body detection in rail transit, and in particular to a method, device, equipment and storage medium for detecting foreign bodies in a railway track area. Background Art

[0002] Urban rail lines, especially ground and elevated ones, face multiple risks from natural and human factors. These risks include natural emergencies such as falling rocks, fallen trees, mudslides, and human-caused risks such as falling objects and illegal intrusions. Therefore, line foreign body monitoring and safety monitoring have become a vital part of the urban rail transit system.

[0003] Traditional foreign object detection methods usually rely on the comparison of the overall point cloud data of the previous and next frames to detect newly appeared foreign objects. Although these methods are effective to a certain extent, they also have obvious limitations, such as sensitivity to environmental changes, neglect of the characteristics of the track area, and low data processing efficiency. The sensitivity to environmental changes makes traditional methods unstable in practical applications and cannot meet the requirements of rail transit systems for high reliability and high stability. The lack of in-depth analysis of the characteristics of the track area leads to a significant reduction in detection accuracy in some complex scenarios, such as when there are interference objects near the track or the background is complex, and even the real foreign objects cannot be effectively identified. The low data processing efficiency not only affects the real-time performance of foreign object detection, but also increases the computing and storage costs of the system, making it difficult to adapt to the needs of large-scale rail transit systems.

[0004] Therefore, how to solve the problems in the existing technology of track foreign object detection, such as being easily affected by environmental changes, lack of in-depth analysis of track area characteristics, and low efficiency of point cloud data processing, has become an important issue in current research. Summary of the invention

[0005] The main technical problem solved by the present invention is to provide a method, device, equipment and storage medium for detecting foreign objects in railway track areas, so as to solve the problems in the prior art that track foreign object detection is easily affected by environmental changes, lacks in-depth analysis of track area characteristics, and has low efficiency in point cloud data processing.

[0006] In order to solve the above technical problems, a technical solution adopted by the present invention is to provide a method for detecting foreign objects in a railway track area, comprising the steps of: performing point cloud segmentation on the point cloud data of the railway environment to obtain a foreign object label set and a track label set; extracting the point cloud data of the track area according to the track label set; comparing the point cloud data of the track area in the previous and next frames to obtain a newly added point cloud data set; performing cluster analysis on the newly added point cloud data set to obtain a foreign object cluster; cross-calculating the foreign object cluster and the foreign object label set to calculate and identify foreign objects in the track area.

[0007] In some embodiments, before performing point cloud segmentation on the point cloud data of the railway environment to obtain the foreign object label set and the track label set step, it also includes: preprocessing the point cloud data of the railway environment, and the preprocessing includes: voxel downsampling and removing outliers in the point cloud data of the railway environment.

[0008] In some embodiments, point cloud data of the track area is extracted according to the track label set. , the method is:

[0009] ;

[0010] in, A label representing a track region, represents the predicted label, Represents point cloud data.

[0011] In some embodiments, the point cloud data of the track area of ​​the previous and next frames are compared to obtain a new point cloud data set. The calculation method is:

[0012] ;

[0013] in, Represents the point cloud data of the track area of ​​the current frame, Point cloud data representing the track area of ​​the previous frame.

[0014] In some embodiments, cluster analysis is performed on the newly added point cloud data set to obtain foreign body clusters by:

[0015] ;

[0016] in, represents the number of data points, represents the number of clusters, Indicates The data points in the clusters, Indicates The centroid of the cluster.

[0017] In some embodiments, the foreign body cluster set and the foreign body label set are cross-calculated by:

[0018] ;

[0019] in, Indicates foreign body clusters, Represents a foreign body tag set.

[0020] In some embodiments, it also includes: when the size of a foreign object cluster in the foreign object cluster set exceeds a first threshold, and the height of the foreign object cluster relative to the rail surface exceeds a second threshold, an alarm signal is generated.

[0021] In order to solve the above technical problems, the present application also provides a railway track area foreign body detection device, comprising:

[0022] The point cloud segmentation module is used to perform point cloud segmentation on the point cloud data of the railway environment to obtain a foreign object label set and a track label set.

[0023] The extraction module is used to extract the point cloud data of the track area according to the track label set.

[0024] The comparison module is used to compare the point cloud data of the track area in the previous and next frames to obtain a new point cloud data set.

[0025] The cluster analysis module is used to perform cluster analysis on the newly added point cloud data set to obtain foreign body clusters.

[0026] The calculation and identification module is used to cross-calculate the foreign object clusters and the foreign object label sets to calculate and identify foreign objects in the track area.

[0027] In order to solve the above technical problems, the present application also provides an electronic device, including a processor and a memory, wherein the memory stores a computer program that can be executed by the processor, and when the computer program is executed by the processor, any one of the above-mentioned methods for detecting foreign objects in railway track areas is implemented.

[0028] In order to solve the above technical problems, the present application also provides a computer-readable storage medium, on which a computer program is stored, and the computer program is used to implement any of the above-mentioned methods for detecting foreign objects in a railway track area.

[0029] Beneficial effects: The present invention discloses a method, device, equipment and storage medium for detecting foreign objects in a railway track area, wherein the method comprises the steps of: performing point cloud segmentation on the point cloud data of the railway environment to obtain a foreign object label set and a track label set; extracting point cloud data of the track area according to the track label set; comparing the point cloud data of the track area of ​​the previous and next frames to obtain a newly added point cloud data set; performing cluster analysis on the newly added point cloud data set to obtain a foreign object cluster; cross-calculating the foreign object cluster with the foreign object label set to calculate and identify foreign objects in the track area. The method of the present application accurately identifies and extracts point cloud data of the track area through point cloud segmentation, and identifies foreign objects in the track area by cross-calculating the foreign object cluster with the foreign object label set, which can not only reduce the interference of environmental changes on foreign object detection, but also improve detection accuracy and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1A flow chart of an embodiment of a method for detecting foreign objects in a railway track area of ​​the present application is provided;

[0031] Figure 2 An example scene diagram of the collected railway environment is given;

[0032] Figure 3 Given the basis Figure 2 The predicted label map obtained;

[0033] Figure 4 Given the basis Figure 3 Point cloud data map of the extracted track area;

[0034] Figure 5 An example of a foreign body cluster obtained according to the method of the present application is given;

[0035] Figure 6 Another example of a foreign body cluster obtained by the method of the present application is given;

[0036] Figure 7 Another example of a foreign body cluster obtained by the method of the present application is given;

[0037] Figure 8 Another example of a foreign body cluster obtained by the method of the present application is given;

[0038] Fig. 9 A schematic structural diagram of an embodiment of a railway track area foreign body detection device of the present application is given;

[0039] Fig.10 A schematic diagram of the framework of an electronic device of the present application is provided;

[0040] Fig.11 A schematic diagram of a computer-readable storage medium according to an embodiment of the present application is provided. DETAILED DESCRIPTION

[0041] In order to facilitate the understanding of the present invention, the present invention is described in more detail below in conjunction with the accompanying drawings and specific embodiments. Preferred embodiments of the present invention are provided in the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described in this specification. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present invention more thorough and comprehensive.

[0042] It should be noted that, unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used in this specification includes any and all combinations of one or more related listed items.

[0043] It should be noted that the "track area" referred to in this application refers to the track and the three-dimensional space within a certain range near the track, including the ground space, the space within a certain range above the track, and the underground space below the track. The "point cloud" referred to in this application refers to the data set of all measurement points of the measurement target in the same spatial reference coordinate system, including the three-dimensional coordinates of each measurement point.

[0044] Figure 1 A flow chart showing an embodiment of a method for detecting foreign objects in a railway track area according to the present invention includes the following steps:

[0045] Step S11, performing point cloud segmentation on the point cloud data of the railway environment to obtain a foreign object label set and a track label set.

[0046] Step S12, extracting point cloud data of the track area according to the track label set.

[0047] Step S13, comparing the point cloud data of the track area in the previous and next frames to obtain a new point cloud data set.

[0048] Step S14: cluster analysis is performed on the newly added point cloud data set to obtain foreign body clusters.

[0049] Step S15, cross-calculating the foreign object clusters and the foreign object label sets to calculate and identify foreign objects in the track area.

[0050] The method of the present application identifies foreign objects in the track area through point cloud segmentation and cross-calculation methods, which can not only accurately identify and extract point cloud data of the track area, but also reduce the interference of environmental changes on foreign object detection, thereby improving detection accuracy and efficiency.

[0051] In some embodiments, before performing point cloud segmentation on the point cloud data of the railway environment to obtain the foreign object label set and the track label set step, it also includes: preprocessing the point cloud data of the railway environment, and the preprocessing includes: voxel downsampling and removing outliers in the point cloud data of the railway environment.

[0052] For example, laser radar equipment can be used to collect point cloud data of the railway environment in real time to generate a model of the track and the surrounding environment in three-dimensional space. Since laser radar equipment is not restricted by conditions such as light and weather, it can still collect point cloud data even in harsh environments such as night and fog, so it can provide a stable detection effect for foreign object detection. Furthermore, the collected point cloud data is downsampled by voxel to reduce the amount of calculation and maintain sufficient spatial resolution; further, by counting the density of points in the neighborhood and removing outliers or noise points, the quality and consistency of the point cloud data in the track area are ensured, thereby enhancing the robustness of the system.

[0053] Among them, the voxel downsampling method is: first grid the point cloud space, also known as voxelization, each grid after gridding is called a voxel, and these extremely small grids contain some points, and then these points are averaged or weighted averaged to get a point to replace all the points in the original grid. Outliers refer to a small number of data points that are abnormal from the normal data set. Noise points refer to data points that have errors or anomalies (deviates from the expected value) in the data, which interfere with the analysis of the data.

[0054] Specifically, assume that the point cloud data of the railway environment collected is ,in, , ,..., Respectively represent the point cloud information of each measurement point in the railway environment, represents the total number of measurement points in the railway environment; , ,in, , , Respectively represent Measurement points The three-dimensional coordinates of x Axis coordinates, y Axis coordinates, z Axis coordinates) information. Further, set the voxel grid size to v , each point in the grid is represented by a point, so that Simplified to ,in, Represents the number of data points in the point cloud data after voxel downsampling, .

[0055] In some embodiments, the point cloud data of the railway environment is segmented to obtain a foreign object label set and a track label set, and the method is: using a neural network model to predict the label of each point in the point cloud data; extracting the labels of the track area and the foreign object points to obtain a foreign object label set and a track label set. Extracting the point cloud data of the track area through the predicted track label set can avoid the noise interference in the overall scene comparison of traditional technologies, and can also effectively improve the accuracy of foreign object detection.

[0056] Exemplarily, the above-mentioned neural network model can be a network model based on Unet. Unet is a convolutional neural network based on deep learning, which can be used for point cloud segmentation tasks, perform semantic segmentation on the input point cloud data, determine the category of each point in the point cloud data and perform precise segmentation.

[0057] Specifically, the preprocessed point cloud data is supervised and trained through a neural network model to predict the label of each observation point, which includes the labels of track points, background points, and possible foreign body points. Assume that is the predicted label set, , ,..., All represent predicted labels, where ( )for .

[0058] In some embodiments, point cloud segmentation of point cloud data of a railway environment is an iterative image segmentation method that starts from one or more "seed points" and expands the area by checking whether adjacent seed points meet a preset similarity criterion. The specific method includes:

[0059] First, one or more initial seed points are selected. These seed points are usually located at the center of the target area and can be selected based on the geometric characteristics of the seed points (such as curvature, normal vector, distance).

[0060] Then, a similarity metric (such as distance, normal vector angle) is determined to determine whether the new seed point can be added to the current expanded area.

[0061] Next, starting from the initial seed point, check whether the adjacent seed points meet the similarity criteria. If so, add the seed point to the region and continue to expand outward.

[0062] The above process is continued until no new seed points meet the similarity criteria. After the region growth is completed, the expanded area is marked as a segmentation area, and the point cloud data in the area is the specific part to be extracted.

[0063] In some embodiments, point cloud segmentation can also be performed using a segmentation method based on plane extraction. Specifically, the segmentation method based on plane extraction uses the geometric features of points in the point cloud to achieve segmentation by detecting and extracting planes in the point cloud:

[0064] First, the main planes in the point cloud can be detected by the RANSAC (Random Sampling Consensus) algorithm. The RANSAC algorithm fits a plane model by randomly selecting several points.

[0065] Then, other points are verified to belong to the plane to find the best plane fit.

[0066] Next, after the plane is detected, plane segmentation is performed, usually taking the point with a high degree of fit to the plane as a point on the plane and separating the point from other points.

[0067] Finally, the fitted plane points are extracted from the point cloud as an independent segmented region, which usually represents areas such as tracks and tunnel walls.

[0068] In some embodiments, point cloud data of the track area is extracted according to the track label set. , the method is:

[0069] ;

[0070] in, A label representing a track region, represents the predicted label, Represents point cloud data.

[0071] In order to further explain the content of this application, Figure 2 An example scene graph of the collected railway environment is given. The example scene graph consists of point cloud data of the railway scene collected in real time. Figure 2 The railway environment in the figure includes the track area, tunnel entrance and other tunnel backgrounds. In order to further extract the point cloud data of the track area, it is necessary to first segment the point cloud data of the above railway scene and use the neural network model to predict the label of each pixel in the example scene image. The results are as follows: Figure 3 As shown, different colors are used to represent different predicted labels to help understanding. Figure 3 The blue point 11 in the figure can represent the predicted label of the track area, the green point 12 represents the predicted label of the tunnel entrance, the yellow point 13 represents the predicted label of the tunnel entrance contact network, and the red point 14 represents the predicted label of other tunnel backgrounds. Further, the point cloud data of the track area is extracted according to the predicted track label set. The extracted results are as follows: Figure 4 As shown, Figure 4 The area consisting of blue points 11 in FIG. 1 represents the point cloud data of the track area.

[0072] In some embodiments, the point cloud data of the track area of ​​the previous and next frames are compared to obtain a new point cloud data set, and the method further includes: subtracting the point cloud data of the background points from the new point cloud data set. Since the method of the present application only compares the point cloud data of the track area, rather than the entire railway environment, unnecessary data processing is reduced, the efficiency and accuracy of foreign object detection are effectively improved, and the probability of false alarms and missed alarms is reduced.

[0073] In some embodiments, a new point cloud dataset is added The calculation method is:

[0074] ;

[0075] in, Represents the point cloud data of the track area of ​​the current frame, Point cloud data representing the track area of ​​the previous frame.

[0076] In some embodiments, cluster analysis is performed on the newly added point cloud data set to obtain foreign body clusters by:

[0077] ;

[0078] in, represents the number of data points, represents the number of clusters, Indicates The data points in the clusters, Indicates The centroid of the cluster.

[0079] Furthermore, the foreign body clusters obtained through cluster analysis can be expressed as ,in, , ,..., They represent foreign body clusters respectively.

[0080] In some embodiments, the foreign body cluster set and the foreign body label set are cross-calculated, and the calculation method is:

[0081] ;

[0082] in, Indicates foreign body clusters, Represents a foreign body tag set.

[0083] In some embodiments, the method further includes: when the size of the foreign matter cluster in the foreign matter cluster exceeds a first threshold value, and the height of the foreign matter cluster relative to the rail surface exceeds a second threshold value, that is, when and When an alarm occurs, it generates an alarm signal to remind railway staff to ensure the safety of railway operation. represents the first threshold, Indicates foreign body cluster Height relative to the rail surface, Indicates the second threshold.

[0084] Specifically, we can first filter according to the size of the foreign body clusters, filter out small foreign body clusters that do not affect driving safety, such as rain noise or track debris, and focus on obstacles that pose a greater threat to track driving safety. The filtering method is:

[0085] ;

[0086] in, Indicates foreign body cluster The size, Indicates the size threshold.

[0087] In some embodiments, the foreign object clusters in the foreign object cluster set can also be merged to simplify the scene and focus on important obstacles. The merging condition is: the distance measurement between the two foreign object clusters is less than the distance threshold. That is: ,in, Indicates foreign body cluster Clusters of foreign matter The distance measure between Indicates the distance threshold.

[0088] Figures 5 to 8 Examples of different foreign body clusters are given, Figure 5 The type of foreign body cluster is stationary person 15, Figure 6 The type of foreign object cluster is moving pedestrian 16, Figure 7 The type of foreign body cluster is rockfall 17, Figure 8 The type of foreign body cluster is train 18. Figures 5 to 8 It can be seen that the foreign object cluster obtained according to the railway track area foreign object detection method of the present application only contains information about foreign objects, so it is easy to identify foreign objects appearing in the track area, and can effectively improve the efficiency and accuracy of foreign object detection.

[0089] In some embodiments, high-resolution cameras or multispectral cameras can also be used to monitor the railway environment and identify foreign objects through computer vision technologies such as image segmentation, object detection, and motion detection.

[0090] In some embodiments, millimeter wave radar detection can also be used for foreign object detection. Millimeter wave radar detection detects three-dimensional information of surrounding objects through electromagnetic waves and has strong robustness to bad weather and light conditions. Millimeter wave radar detection can also be used in combination with deep learning models to process the signals returned by the radar, thereby detecting and classifying foreign objects on the railway track area.

[0091] In some embodiments, multiple sensors such as radar, camera and millimeter wave radar can be combined for foreign object detection, and the complementary characteristics of different sensors can be used to enhance the robustness and accuracy of foreign object detection. Multi-sensor fusion can integrate data from different sensors through data fusion algorithms to achieve higher detection accuracy and stronger anti-interference capabilities in track area and foreign object detection.

[0092] Combination Fig. 9 The present application also provides a railway track area foreign body detection device, comprising:

[0093] The point cloud segmentation module 21 is used to perform point cloud segmentation on the point cloud data of the railway environment to obtain a foreign object label set and a track label set.

[0094] The extraction module 22 is used to extract the point cloud data of the track area according to the track label set.

[0095] The comparison module 23 is used to compare the point cloud data of the track area in the previous and next frames to obtain a new point cloud data set.

[0096] The cluster analysis module 24 is used to perform cluster analysis on the newly added point cloud data set to obtain foreign body clusters.

[0097] The calculation and identification module 25 is used to perform cross calculations on the foreign object clusters and the foreign object label sets to calculate and identify foreign objects in the track area.

[0098] The other technical features of the above-mentioned railway track area foreign object detection device are the same as the features disclosed in the above-mentioned embodiment method, and will not be repeated here.

[0099] Combination Fig.10 The present application also provides an electronic device, including a processor 32 and a memory 31, wherein the memory 31 stores a computer program that can be executed by the processor, and when the computer program is executed by the processor 32, any one of the above-mentioned methods for detecting foreign objects in railway track areas is implemented.

[0100] Specifically, the memory 31 is used to store computer programs or data. The memory 31 can be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc.

[0101] The processor 32 is used to read / write data or computer programs stored in the memory 31 and execute corresponding functions. For example, when the computer program stored in the memory 31 is executed by the processor 32, a method for detecting foreign objects in a railway track area provided in an embodiment of the present application can be implemented.

[0102] Combination Fig.11 The present application also provides a computer-readable storage medium 40 on which a computer program 401 is stored. The computer program 401 is used to implement any of the above-mentioned methods for detecting foreign objects in a railway track area.

[0103] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0104] The above description of various embodiments tends to emphasize the differences between the various embodiments. The same or similar aspects can be referenced to each other, and for the sake of brevity, they will not be repeated herein.

[0105] In the several embodiments provided in the present application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device implementation described above is only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation, such as 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 interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0106] 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. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0107] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or 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, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of each implementation method of the present application. The aforementioned storage medium can be non-transient or transient. Specifically, it includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), disk or optical disk and other media that can store program code.

[0108] In summary, the present invention discloses a method, device, equipment and storage medium for detecting foreign objects in a railway track area, wherein the method includes the steps of: performing point cloud segmentation on the point cloud data of the railway environment to obtain a foreign object label set and a track label set; extracting point cloud data of the track area according to the track label set; comparing the point cloud data of the track area of ​​the previous and next frames to obtain a newly added point cloud data set; performing cluster analysis on the newly added point cloud data set to obtain a foreign object cluster; cross-calculating the foreign object cluster with the foreign object label set to calculate and identify foreign objects in the track area. The method of the present application accurately identifies and extracts point cloud data of the track area through point cloud segmentation, and identifies foreign objects in the track area by cross-calculating the foreign object cluster with the foreign object label set, which can not only reduce the interference of environmental changes on foreign object detection, but also improve detection accuracy and efficiency.

[0109] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention's specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for detecting foreign objects in a railway track area, characterized in that: Includes steps: Point cloud segmentation is performed on the point cloud data of the railway environment to obtain a foreign object label set and a track label set, and the method is: using a neural network model to predict the label of each point in the point cloud data; extracting the labels of the track area and the foreign object points to obtain the foreign object label set and the track label set; wherein the neural network model is used to perform supervised training on the pre-processed point cloud data to predict the label of each observation point, which includes the labels of the track point, background point and possible foreign object point. To predict the label set, , ,..., All represent predicted labels, where Point cloud data The predicted label of ; The prediction labels include prediction labels for track areas, prediction labels for tunnel entrances, prediction labels for tunnel entrance contact networks, and prediction labels for other tunnel backgrounds; Extracting point cloud data of the track area according to the track label set; Comparing the point cloud data of the track area in the previous and next frames to obtain a new point cloud data set; further comprising: subtracting the point cloud data of the background points from the new point cloud data set; Performing cluster analysis on the newly added point cloud data set to obtain foreign body clusters; The foreign object cluster set and the foreign object label set are cross-calculated to calculate and identify foreign objects in the track area.

2. The method for detecting foreign objects in railway track areas according to claim 1, characterized in that: Before the step of performing point cloud segmentation on the point cloud data of the railway environment to obtain a foreign body label set and a track label set, the step further includes: The point cloud data of the railway environment is preprocessed, and the preprocessing includes: voxel downsampling and removing outliers in the point cloud data of the railway environment.

3. The method for detecting foreign objects in railway track areas according to claim 1, characterized in that: Extracting point cloud data of the track area according to the track label set , the method is: ； in, A label representing a track region, represents the predicted label, Represents point cloud data.

4. The method for detecting foreign objects in railway track areas according to claim 1, characterized in that: The point cloud data of the track area of ​​the previous and next frames are compared to obtain a new point cloud data set. The calculation method is: ； in, Represents the point cloud data of the track area of ​​the current frame, Point cloud data representing the track area of ​​the previous frame.

5. The method for detecting foreign objects in railway track areas according to claim 1, characterized in that: The method of clustering the newly added point cloud data set to obtain foreign body clusters is as follows: ； in, represents the number of data points, represents the number of clusters, Indicates The data points in the clusters, Indicates The centroid of the cluster.

6. The method for detecting foreign objects in railway track areas according to claim 1, characterized in that: The cross calculation of the foreign body cluster set and the foreign body label set is performed by: ； in, Indicates foreign body clusters, Represents a foreign body tag set.

7. The method for detecting foreign objects in railway track areas according to claim 1, characterized in that: Also includes: When the size of the foreign matter cluster in the foreign matter cluster set exceeds a first threshold, and the height of the foreign matter cluster relative to the rail surface exceeds a second threshold, an alarm signal is generated.

8. A foreign body detection device for railway track area, characterized in that: include: The point cloud segmentation module is used to perform point cloud segmentation on the point cloud data of the railway environment to obtain a foreign object label set and a track label set, including: using a neural network model to predict the label of each point in the point cloud data; extracting the labels of the track area and the foreign object points to obtain the foreign object label set and the track label set; wherein the neural network model is used to perform supervised training on the pre-processed point cloud data to predict the label of each observation point, which includes the labels of the track point, background point and possible foreign object point. To predict the label set, , ,..., All represent predicted labels, where Point cloud data The predicted label of ;;The prediction labels include prediction labels for track areas, prediction labels for tunnel entrances, prediction labels for tunnel entrance contact networks, and prediction labels for other tunnel backgrounds; An extraction module, used for extracting point cloud data of the track area according to the track label set; A comparison module is used to compare the point cloud data of the track area of ​​the previous and next frames to obtain a new point cloud data set; and further includes: subtracting the point cloud data of the background points from the new point cloud data set; A cluster analysis module, used for performing cluster analysis on the newly added point cloud data set to obtain foreign body clusters; The calculation and identification module is used to perform cross calculation on the foreign object cluster set and the foreign object label set to calculate and identify foreign objects in the track area.

9. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory stores a computer program that can be executed by the processor, and when the computer program is executed by the processor, the method for detecting foreign objects in a railway track area according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is used to implement the railway track area foreign object detection method as described in any one of claims 1 to 7.

Images