Railway track area block drop detection method and system

Abstract

The present invention discloses a method for detecting block drop in a railway track area, comprising a track area image acquisition step, in which a visual camera is used to collect an image of a running track area as a sleeper image; a block drop image detection step, in which block drop features in the sleeper image are extracted based on a preset image edge detection algorithm model, wherein the image edge detection algorithm model is configured to fuse Canny edge detection and multi-scale gradient feature extraction; a block drop verification step, in which, when a block drop feature is detected, a block drop verification device is started to spray an air flow toward an area where the block drop feature is located to drive the block drop to move; an image continuous acquisition step, in which, after the block drop verification device is started, track area images are continuously acquired as verification images; a block drop position comparison and output step, in which the sleeper image and the verification image are aligned, a coordinate system is divided based on two tracks, and the relative position offset of the block drop feature in the two track coordinate systems is compared. If the offset exceeds a preset threshold, it is determined to be a real block drop and a marking result is output.

Description

Technical Field

[0001] The present invention relates to the field of railway track detection, and in particular to a method and system for detecting block drop in a railway track area. Background Art

[0002] With the rapid development of my country's rail transit industry, the mileage of rail laying has increased year by year. In rail transit, the repeated rolling of train wheels can cause various hazards to the rail surface, such as wear, joint collapse, block falling and track cracking. Among them, the problem of block falling is particularly serious. If the location of the fallen block cannot be discovered and repaired in time, the gap formed will lead to increased wheel impact and even cause derailment accidents. In addition, delayed treatment may also lead to the need to replace the entire section of rail, the cost of which is much higher than local repair.

[0003] At present, the main methods for detecting fallen blocks include manual inspections and the use of track inspection vehicles. Track inspection vehicles are usually equipped with high-definition high-speed cameras, laser scanners or ultrasonic flaw detectors to detect fallen blocks or the location of fallen blocks in the track area. Among them, high-definition high-speed cameras continuously take pictures of the track area and use image processing algorithms to detect whether there are fallen blocks. This method mainly relies on identifying objects in the image that are different in color from the sleepers and have regular shapes, and marking them as potential fallen blocks.

[0004] Existing chip drop detection technology has some obvious limitations: First, in areas with complex terrain, such as subway systems in mountainous or coastal areas, tracks may switch between underground and ground. In underground environments, due to fewer interference factors, the image algorithm has a higher accuracy rate in identifying chip drops. However, on the ground, the environment is more complex, and rainwater, bird droppings, etc. may fall into the track area, causing serious interference to the recognition of the image algorithm, making the identification of chip drops on open-air tracks difficult. Secondly, existing image recognition algorithms mainly rely on color differences and shape regularity to judge chip drops. This method is easily affected by environmental factors such as lighting changes, shadows, stains, etc., resulting in misjudgments or missed judgments. Relying solely on image recognition, it is impossible to determine whether the detected anomaly is a real chip drop or other foreign matter, which may lead to unnecessary manual inspections and increase maintenance costs and workload. In addition, existing technologies lack the ability to quantitatively analyze chip drops and cannot accurately estimate the size, location and potential harm of chip drops, which is crucial for formulating effective maintenance strategies.

[0005] The existing detection system lacks a real-time verification mechanism. Once suspected block loss is found, manual inspections are usually required for confirmation. This is not only time-consuming and labor-intensive, but may also delay the timely handling of actual block loss. Summary of the Invention

[0006] The purpose of this application is to provide a method and system for detecting block drop in railway track areas, which has the advantages of improving the accuracy of block drop detection, reducing misjudgments and missed judgments, realizing real-time verification, and providing quantitative analysis capabilities.

[0007] The present application provides a method for detecting block drop in a railway track area, and the technical solution is as follows: It comprises the following steps: a track area image acquisition step, in which a track area image of a train is collected as a sleeper image by a visual camera; a block drop image detection step, in which block drop features in the sleeper image are extracted based on a preset image edge detection algorithm model, and the image edge detection algorithm model is configured to fuse Canny edge detection and multi-scale gradient feature extraction; a block drop verification step, in which when the block drop feature is detected, a block drop verification device is started, and the block drop verification device includes an adjustable jet group for spraying airflow to the area where the block drop feature is located to drive the block drop to move; an image continuous acquisition step, in which after the block drop verification device is started, track area images are continuously acquired as verification images; a block drop position comparison and output step, in which the sleeper image and the verification image are aligned, the coordinate system is divided based on the two tracks, and the relative position offset of the block drop feature in the two track coordinate systems is compared. If the offset exceeds a preset threshold, it is determined to be a real block drop and a marking result is output.

[0008] Furthermore, the present application also proposes that the sleeper image and verification image are both processed by a partitioning algorithm, and are divided into several parallel partitions with two tracks as boundaries. Each partition is parallel to the track direction and has an adjustable width, and each partition corresponds to an independent position code; when the block drop feature is detected to be located in a certain partition, the jet group of the block drop verification device is directional adjusted to match the spatial coordinates of the partition.

[0009] Furthermore, the present application also proposes that the block drop verification device includes at least two jet groups, which are respectively deployed on both sides of the track and symmetrically distributed on both sides of the track center line; each jet group includes a jet nozzle, a swinging component for driving the jet nozzle deflection angle and an air source for providing high-pressure gas to the jet nozzle, and the block drop verification device is equipped with an air path control system; when the block drop feature is located in a partition on one side of the track center line, the jet group close to that side is started, and the swinging component adjusts the jet angle of the jet nozzle according to the partition coordinates, and the jet nozzle drives the airflow to act on the block drop feature area.

[0010] Furthermore, the present application also proposes that the partition width of the verification image is strictly equal to the partition width of the sleeper image, and the partition boundaries are aligned through an image registration algorithm; in the block drop position comparison output step, if the offset of the block drop feature in the verification image relative to the sleeper image spans at least one partition, it is judged as a valid movement, and a block drop motion trajectory report is generated in combination with the air pressure parameters of the jet group and the partition offset distance.

[0011] Furthermore, the present application also proposes that it also includes a block drop feature image verification step, specifically: edge contour extraction of the block drop features in the sleeper image and the verification image respectively, and calculation of the contour similarity and area change rate of the two; if the contour similarity is lower than the preset value and the area change rate does not exceed the threshold, it is determined to be a static foreign body rather than a block drop; if the contour similarity meets the requirements and the area change rate is associated with the position offset, it is determined to be a real block drop.

[0012] Furthermore, the present application also proposes that it also includes the steps of storing and classifying block drop data, specifically: establishing a block drop feature database to store the edge contours, motion trajectories, air pressure parameters and partition offset data of real block drops; clustering analysis of block drop contour features through convolutional neural networks to generate block drop type labels classified by shape and size, and matching them with historical data to optimize the judgment threshold.

[0013] Furthermore, the present application also proposes that it also includes a secondary verification step for dropped blocks, specifically: after the jet verification execution module in the first dropped block verification step is started, if the dropped block feature of the verification image does not move, the secondary verification mode is triggered; in the secondary verification mode, the air path control system increases the air pressure of the jet group to a preset high-pressure threshold, and adjusts the jet nozzle angle through the swing component for secondary injection; after the secondary injection, a secondary verification image is obtained through the image continuous acquisition step, and if the dropped block feature still does not move, it is determined to be a static foreign object, and the feature is marked as a non-dropped block type; and a foreign object database is established to store the position, air pressure parameters and image features of static foreign objects, and is synchronously updated to the railway operation and maintenance management system to eliminate false detection interference.

[0014] Furthermore, the present application also proposes that the block drop image detection step also includes calculating the pixel area, aspect ratio and geometric center coordinates of the block drop feature in the image, and performing spatial mapping in combination with the actual size of the track to output the physical size estimate of the block drop and the coordinates of the block drop position.

[0015] Furthermore, the present application also proposes that the block drop image detection step is configured with a dynamic air pressure adaptation algorithm, specifically including: a block drop area quantification sub-step, in which the pixel area S of the block drop feature in the image is calculated by the edge detection processing module, and converted into a physical area S1 based on the mapping relationship between the actual track size and the image resolution; an air pressure calculation sub-step, in which a basic air pressure threshold P0 is set, and the first block drop verification injection air pressure value P is dynamically adjusted according to the following formula:

[0016] ,

[0017] Among them, S2 is the preset reference area threshold, is the area weight coefficient, and its value range is 0.2 ≤ α ≤ 0.5; in the secondary verification mode, the secondary injection pressure is 1.2 to 2 times of p.

[0018] Furthermore, the present application also proposes a railway track area block drop detection system, including: a track image acquisition module, used to obtain sleeper images and verification images through a visual camera; an edge detection processing module, configured to run an image edge detection algorithm model to extract block drop features; an air jet verification execution module, including a symmetrically arranged air jet group, a swing drive unit and an air pressure control unit, used to directionally drive the block drop movement; an image analysis and comparison module, used to align images, divide partitions and calculate position offsets; a data storage and output module, used to store block drop feature data and generate a detection report.

[0019] From the above, it can be seen that the present application provides a method and system for detecting block drop in the railway track area, which includes the steps of track area image acquisition, block drop image detection, block drop verification, continuous image acquisition and block drop position comparison output. By combining image processing technology and physical verification methods, accurate identification and verification of block drop are achieved, effectively solving the problems of misjudgment and missed judgment in the existing technology, improving the accuracy of block drop detection, and being able to verify in real time and provide quantitative analysis. It has the advantages of improving the accuracy of block drop detection, reducing misjudgment and missed judgment, realizing real-time verification, and providing quantitative analysis capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is an overall flow chart of a method for detecting block drop in a railway track area;

[0021] Figure 2 is the sleeper image of the normal railway track area in the present invention;

[0022] Figure 3 It is the sleeper image containing the block drop feature in the present invention. DETAILED DESCRIPTION

[0023] The present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0024] Example:

[0025] The present application provides a method for detecting block drop in a railway track area, comprising the following steps: combining Figure 1, track area image acquisition step, collecting the track area image of the train as the sleeper image through the visual camera; block drop image detection step, extracting the block drop features in the sleeper image based on the preset image edge detection algorithm model, the image edge detection algorithm model is configured to integrate Canny edge detection and multi-scale gradient feature extraction; block drop verification step, when the block drop feature is detected, starting the block drop verification device, the block drop verification device includes an adjustable jet group, which is used to spray airflow to the area where the block drop feature is located to drive the block drop to move; image continuous acquisition step, after the block drop verification device is started, continuously collecting track area images as verification images; block drop position comparison and output step, aligning the sleeper image and the verification image, dividing the coordinate system based on the two tracks, and comparing the relative position offset of the block drop feature in the two track coordinate systems. If the offset exceeds the preset threshold, it is determined to be a real block drop and the marking result of the real block drop is output.

[0026] In the daily maintenance of railway tracks, the detection and repair of fallen blocks is an important link. If fallen blocks are not repaired in time, it will cause wheel impact during train operation and even cause derailment accidents. Existing detection methods mainly rely on manual inspections and track inspection vehicles. However, in areas with complex terrain, especially open-air tracks, the accuracy and efficiency of detection are greatly affected by environmental interference. This application proposes a fallen block detection method based on a visual camera and an image edge detection algorithm. Through multi-step image processing and verification, the accuracy and reliability of fallen block detection can be effectively improved.

[0027] In the track area image acquisition step, the image of the track area where the train is traveling is collected by a visual camera as the sleeper image. In the block drop image detection step, the block drop features in the sleeper image are extracted based on a preset image edge detection algorithm model. The image edge detection algorithm model integrates Canny edge detection and multi-scale gradient feature extraction. In the block drop verification step, when the block drop feature is detected, the block drop verification device is started. The block drop verification device includes an adjustable jet group for spraying airflow to the area where the block drop feature is located to drive the block drop to move. In the image continuous acquisition step, after the block drop verification device is started, the track area image is continuously collected as the verification image. In the block drop position comparison and output step, the sleeper image and the verification image are aligned, the coordinate system is divided based on the two tracks, and the relative position offset of the block drop feature in the two track coordinate systems is compared. If the offset exceeds the preset threshold, it is determined to be a real block drop and the marking result of the real block drop is output.

[0028] The track area image is collected by a visual camera, and the features of the fallen blocks are extracted based on the image edge detection algorithm model. The fallen block verification device drives the fallen blocks to move by jetting air, and verifies the actual movement of the fallen blocks by continuously collecting images. By comparing the fallen block feature positions in the sleeper image and the verification image, it is determined whether the fallen blocks actually exist. This method can effectively improve the accuracy of fallen block detection and reduce the impact of environmental interference.

[0029] Compared with the existing technology, the method of the present application can more accurately detect the block drop characteristics of the track area through multi-step image processing and verification. The existing manual inspection and track inspection vehicles have low detection accuracy in complex environments, while the method of the present application can effectively reduce environmental interference and improve detection reliability through image edge detection and jet verification.

[0030] This application uses a visual camera to capture images of the track area and extracts the block drop features based on the image edge detection algorithm model. When the block drop features are detected, the block drop verification device is started, and an air flow is sprayed to the area where the block drop features are located to drive the block drop to move. Subsequently, the track area images are continuously captured as verification images. By comparing the block drop feature positions in the sleeper image and the verification image, it is determined whether the block drop actually exists. Through this method, the accuracy and reliability of block drop detection can be effectively improved, and the impact of environmental interference can be reduced.

[0031] Furthermore, the present application also proposes that, in combination with Figure 2 Both the sleeper image and the verification image are processed by a partitioning algorithm, and are divided into several parallel partitions with the two tracks as the boundary. Each partition is parallel to the track direction and has an adjustable width. Each partition corresponds to an independent position code; when the block drop feature is detected to be located in a certain partition, the jet group of the block drop verification device is oriented and adjusted to match the spatial coordinates of the partition.

[0032] In the context of this application, the partitioning processing of the sleeper image and the verification image is to improve the accuracy and efficiency of the block drop detection. By dividing the track image into several parallel partitions, the width of each partition can be adjusted as needed, and each partition has an independent position code. This partitioning processing method makes the detection and positioning of the block drop features more accurate.

[0033] In the technical solution of the present application, the jet group of the block drop verification device can perform directional adjustment according to the partition where the detected block drop feature is located. Specifically, when the block drop feature is detected to be located in a certain partition, the jet group will adjust its jet direction to match the spatial coordinates of the partition. This can ensure that the jet device can accurately align with the location of the block drop feature for jet verification.

[0034] Through the above-mentioned partitioning algorithm processing, the block drop verification device can more accurately locate and verify the block drop characteristics, avoid errors that may occur during the jet verification process, and improve the reliability and accuracy of block drop detection.

[0035] Compared with the existing technology, the technical solution of the present application processes the sleeper image and verification image through a partitioning algorithm, so that the positioning of the block drop features is more accurate, and the jet group can be directional adjusted according to the partition coordinates, thereby improving the accuracy and efficiency of the block drop verification. As a result, the block drop features of the track area can be more effectively detected and verified, avoiding the problems of false detection and missed detection caused by inaccurate block drop detection.

[0036] Furthermore, the present application also proposes that the block drop verification device includes at least two jet groups, which are respectively deployed on both sides of the track and symmetrically distributed on both sides of the track center line; each jet group includes a jet nozzle, a swinging component for driving the jet nozzle deflection angle and an air source for providing high-pressure gas to the jet nozzle, and the block drop verification device is equipped with an air path control system; when the block drop feature is located in a partition on one side of the track center line, the jet group close to that side is started, the swinging component adjusts the jet angle of the jet nozzle according to the partition coordinates, and the jet nozzle drives the airflow to act on the block drop feature area.

[0037] The technical solution of the present application can more accurately verify the block drop characteristics by symmetrically deploying the jet groups on both sides of the track and using the swing component to adjust the jet angle of the jet nozzle. Specifically, when the block drop characteristic is detected to be located in a certain partition, the jet group close to the partition will be started and the angle of the jet nozzle will be adjusted by the swing component so that the airflow can accurately act on the block drop characteristic area. In this way, the block drop is moved by the jet drive, thereby verifying whether it is a real block drop.

[0038] Possible implementations of the new technical features include: 1. The air nozzle can be made of a high-pressure and corrosion-resistant alloy to ensure it is not easily damaged by high-pressure gas. 2. The swing assembly can be driven by an electric servo motor to achieve precise angle control. 3. The gas source can be configured as a high-pressure gas cylinder or a compressed air system, and the pressure can be adjusted through the air path control system to accommodate different block drop characteristics and zoning requirements. 4. The air path control system can be integrated with pressure sensors and control valves to achieve real-time monitoring and adjustment of the air pressure. 5. The system can also be configured with redundant air injection groups to improve system reliability and stability.

[0039] The solution of the present application improves the accuracy and reliability of block drop verification by symmetrically arranging the jet groups and precisely controlling the jet nozzle angle. Compared with the existing technology, the present application can more effectively verify the characteristics of block drop, reduce false detection and missed detection, and through precise control of air pressure and jet angle, can adapt to different track environments and changes in block drop characteristics, thereby improving the efficiency and safety of track maintenance.

[0040] Furthermore, the present application also proposes that the partition width of the verification image is strictly equal to the partition width of the sleeper image, and the partition boundaries are aligned through the image registration algorithm. In the block position comparison output step, if the offset of the block feature in the verification image relative to the sleeper image spans at least one partition, it is judged as a valid movement, and the block motion trajectory report is generated in combination with the air pressure parameters of the jet group and the partition offset distance.

[0041] The technical solution of the present application aims to improve the accuracy and reliability of block drop detection. By ensuring that the partition widths of the verification image and the sleeper image are strictly equal, and aligning the partition boundaries through an image registration algorithm, the accuracy in comparing the block drop positions can be ensured. When the position of the block drop feature in the verification image is offset relative to the sleeper image, and the offset exceeds one partition, the system can accurately determine the effective movement of the block drop, and generate a block drop motion trajectory report in combination with the air pressure parameters of the jet group and the partition offset distance to provide more detailed block drop motion information.

[0042] The technical feature of strictly equal partition widths of the verification image and the sleeper image can be achieved in the following ways: first, during image acquisition, a high-precision visual camera is used to ensure consistent image resolution; second, during image processing, a unified partitioning algorithm is used to divide the image to ensure that the width of each partition is consistent; finally, the partition boundaries are aligned through an image registration algorithm to ensure the consistency of partition positions between different images.

[0043] The technical feature of generating a block drop motion trajectory report can be achieved in the following way: when the effective movement of the block drop feature is detected, the motion trajectory of the block drop is calculated in combination with the air pressure parameters of the jet group and the partition offset distance. Specifically, the moving distance of the block drop under the action of the jet can be determined based on the air pressure parameters of the jet group, and the motion trajectory report of the block drop can be generated in combination with the partition offset in the image. The report can provide detailed information such as the moving path and moving distance of the block drop to facilitate further analysis and processing.

[0044] This application improves the accuracy and reliability of block drop detection by strictly aligning the partition width of the verification image and the sleeper image, and combining the air pressure parameters of the jet group and the partition offset distance to generate a block drop movement trajectory report. Compared with the existing technology, this application can more accurately determine the effective movement of the block drop and provide detailed block drop movement information, which helps to timely discover and deal with the block drop problem and ensure the safe operation of rail transit.

[0045] Furthermore, the present application also proposes a step for verifying the image with missing blocks, specifically:

[0046] Combine Figure 3 , the edge contours of the block drop features in the sleeper image and the verification image are extracted respectively, and the contour similarity and area change rate of the two are calculated. If the contour similarity is lower than the preset value and the area change rate does not exceed the threshold, it is judged as a static foreign object rather than a block drop. If the contour similarity meets the requirements and the area change rate is associated with the position offset, it is judged as a real block drop and the real block drop result is output.

[0047] The technical solution of the chip drop feature image verification step is to further improve the accuracy of chip drop detection to avoid misidentifying static foreign objects as chip drops. By extracting the edge contours of the chip drop features in the sleeper image and the verification image, the shape features of the chip drop can be accurately obtained, and the contour similarity and area change rate can be calculated. It can effectively determine whether the chip drop features have changed, thereby distinguishing between real chip drops and static foreign objects.

[0048] Specifically, edge contour extraction can use classic edge detection algorithms, such as Canny edge detection, combined with image processing technology to extract the edge contour of the block feature. Contour similarity calculation can use shape matching algorithms, such as Hausdorff distance or shape context descriptor, to evaluate the similarity of two contours. The area change rate is determined by calculating the area difference between the block feature in the sleeper image and the verification image.

[0049] For example, if a piece of debris is detected in the sleeper image, the piece of debris verification device is started, the verification image is obtained, and the edge contour of the piece of debris in the verification image is extracted. The contour similarity and area change rate of the piece of debris in the sleeper image and the verification image are calculated. If the contour similarity is lower than the preset value (such as 0.8) and the area change rate does not exceed the threshold (such as 10%), the feature is judged to be a static foreign body. If the contour similarity meets the requirements (such as greater than or equal to 0.8) and the area change rate is associated with the position offset, it is judged to be a real piece of debris.

[0050] Therefore, through the block drop feature image verification step, the accuracy of block drop detection can be further improved, the false detection rate can be reduced, and the safety and reliability of railway tracks can be ensured. Compared with the existing technology, the technical solution of this application can more effectively distinguish between block drop and static foreign objects, avoiding resource waste and safety hazards caused by false detection.

[0051] Furthermore, the present application also proposes to establish a block drop feature database to store the edge contours, motion trajectories, air pressure parameters and partition offset data of real block drops; cluster analysis of block drop contour features is performed through convolutional neural networks to generate block drop type labels classified by shape and size, and match them with historical data to optimize the judgment threshold.

[0052] The technical solution of this application mainly involves establishing a block drop feature database and using a convolutional neural network to perform cluster analysis on the block drop contour features. By storing the edge contours, motion trajectories, air pressure parameters and partition offset data of real block drops, the block drop features can be systematically managed and analyzed. The convolutional neural network is used to perform cluster analysis on these features to generate block drop type labels classified by shape and size. By matching with historical data, the threshold for block drop judgment can be continuously optimized to improve the accuracy and reliability of detection.

[0053] There are many ways to establish a database of block drop features. For example, a dedicated data storage system can be designed to store and manage the edge contours, motion trajectories, air pressure parameters and partition offset data of the blocks. The implementation of the convolutional neural network can be based on existing deep learning frameworks such as TensorFlow or PyTorch, and the block drop features can be identified and classified by training the model. The clustering analysis of the data can use common clustering algorithms such as K-means to classify the blocks according to their shape and size. By continuously updating and optimizing the judgment threshold, the detection system can maintain efficient detection performance in different environments and conditions.

[0054] By establishing a block drop feature database and using convolutional neural networks for cluster analysis, the present application can effectively manage and analyze block drop data and improve the accuracy and reliability of block drop detection. Compared with the existing technology, the present application provides a systematic and intelligent block drop detection method that can better adapt to the complex and changeable track environment and reduce the occurrence of false detection and missed detection. Therefore, the present application has important practical significance in improving track safety and maintenance efficiency.

[0055] Furthermore, the present application also proposes a secondary verification step for dropped blocks. When the jet verification execution module is started in the first dropped block verification step, if the dropped block feature of the verification image does not move, the secondary verification mode is triggered. In the secondary verification mode, the air path control system increases the air pressure of the jet group to a preset high-pressure threshold, and adjusts the jet nozzle angle through the swing component for secondary injection. After the secondary injection, the secondary verification image is obtained through the continuous image acquisition step. If the dropped block feature still does not move, it is determined to be a static foreign body, and the feature is marked as a non-dropped block type. A foreign body database is established to record static foreign bodies for storing the position, air pressure parameters and image features of static foreign bodies, and synchronously updated to the railway operation and maintenance management system to eliminate false detection interference.

[0056] The secondary verification step of the fallen block in this application is intended to further improve the accuracy of the fallen block detection and solve the problem of false detection that may exist in the first verification. By increasing the air pressure of the jet group and adjusting the jet nozzle angle for secondary injection, the fallen block can be driven to move more effectively, thereby more accurately judging whether the fallen block feature is a real fallen block.

[0057] Furthermore, the block drop secondary verification step of the present application adopts a higher air pressure and an adjustment of the air nozzle angle to ensure that it can fully act on the block drop feature area during the secondary verification process. If the block drop feature still does not move after the secondary verification, it can be more accurately judged as a static foreign object and the feature can be marked as a non-block drop type. This process not only improves the accuracy of block drop detection, but also further optimizes the performance of the block drop detection system by establishing a foreign object database to store the location, air pressure parameters and image features of static foreign objects.

[0058] In general, this application significantly improves the accuracy and reliability of block drop detection by adding a secondary verification step for block drop. Compared with the existing technology, this application can more effectively eliminate false detection interference, ensure the accuracy of block drop detection results, and thus improve the efficiency and safety of railway track maintenance.

[0059] Furthermore, the present application also proposes to calculate the pixel area, aspect ratio and geometric center coordinates of the block drop feature in the image, and perform spatial mapping based on the actual size of the track to output the estimated physical size of the block drop and the coordinates of the block drop position.

[0060] During the block drop detection process, a visual camera is used to obtain sleeper images and verification images, and the block drop features are extracted through the image edge detection algorithm model. The pixel area, aspect ratio and geometric center coordinates of the block drop features in the image are calculated. Based on the mapping relationship between the actual size of the track and the image resolution, these image features are converted into physical size estimates and the coordinates of the block drop location.

[0061] New technical features include calculating the pixel area, aspect ratio and geometric center coordinates of the block drop feature, and performing spatial mapping based on the actual size of the track. These technical features can be achieved in the following ways: first, extracting the edge contour of the block drop feature through an image processing algorithm, and then calculating the pixel area enclosed by the contour. Secondly, calculating the aspect ratio of the block drop feature and determining its geometric center coordinates. Finally, using the mapping relationship between the actual size of the track and the image resolution, the above image features are converted into physical size estimates and the coordinates of the block drop location.

[0062] Through the above technical solution, the physical size and falling position of the fallen block can be accurately estimated, thereby improving the accuracy and reliability of fallen block detection. Compared with the existing technology, the solution of this application can provide more accurate fallen block positioning and size estimation, which helps to timely discover and repair fallen blocks and avoid safety hazards caused by fallen blocks.

[0063] Furthermore, the present application also proposes a railway track area block drop detection system, comprising:

[0064] The track image acquisition module is used to obtain sleeper images and verification images through a visual camera; the edge detection processing module is configured to run the image edge detection algorithm model to extract the features of the fallen blocks; the jet verification execution module includes a symmetrically arranged jet group, a swing drive unit and an air pressure control unit, which are used to directionally drive the fallen blocks to move; the image analysis and comparison module is used to align images, divide partitions and calculate position offsets; the data storage and output module is used to store the feature data of the fallen blocks and generate inspection reports.

[0065] The technical solution of this system aims to solve the problems of false detection and missed detection in track block detection. The track image acquisition module obtains the sleeper image and verification image. The edge detection processing module processes the image and extracts the block features. The jet verification execution module drives the block to move in a direction through the jet group, swing drive unit and air pressure control unit to verify the authenticity of the block features. The image analysis and comparison module aligns the image, divides the partitions and calculates the position offset of the block features. The data storage and output module is responsible for storing the block feature data and generating a detection report.

[0066] New technical features of the system include a dynamic air pressure adaptation algorithm, which includes: a block drop area quantification sub-step, in which the edge detection processing module calculates the pixel area S of the block drop feature in the image and converts it into a physical area S1 based on the mapping relationship between the actual track size and image resolution; an air pressure calculation sub-step, in which a basic air pressure threshold P0 is set and the initial block drop verification injection pressure value P is dynamically adjusted according to the following formula:

[0067] ,

[0068] Among them, S2 is the preset reference area threshold, is the area weight coefficient, and its value range is 0.2 ≤ α ≤0.5.

[0069] The pixel area of the block drop feature in the image is calculated through the edge detection processing module and converted into a physical area based on the mapping relationship between the actual track size and the image resolution. The air pressure calculation sub-step sets the basic air pressure threshold and dynamically adjusts the first block drop verification injection pressure value according to the formula. In the secondary verification mode, the secondary injection pressure is 1.2 to 2 times the initial injection pressure.

[0070] Furthermore, the system can dynamically adjust the injection air pressure according to the physical area of the block drop features through a dynamic air pressure adaptation algorithm, thereby improving the accuracy and reliability of block drop verification. In this way, the system can more effectively detect and verify the block drop features of the track area, reduce false detections and missed detections, and provide reliable data support for track maintenance.

[0071] Furthermore, the present application also proposes a railway track area block drop detection system, comprising:

[0072] The track image acquisition module is used to obtain sleeper images and verification images through a visual camera; the edge detection processing module is configured to run the image edge detection algorithm model to extract the features of the fallen blocks; the jet verification execution module includes a symmetrically arranged jet group, a swing drive unit and an air pressure control unit, which are used to directionally drive the fallen blocks to move; the image analysis and comparison module is used to align images, divide partitions and calculate position offsets; the data storage and output module is used to store the feature data of the fallen blocks and generate inspection reports.

[0073] The system uses a visual camera to obtain sleeper images and verification images, uses an image edge detection algorithm to extract the features of the fallen blocks, and uses the jet verification execution module to directionally drive the fallen blocks to move. The image analysis and comparison module is responsible for aligning images, dividing partitions and calculating position offsets. The data storage and output module is used to store the fallen block feature data and generate inspection reports.

[0074] Furthermore, the jet verification execution module includes a symmetrically arranged jet group, a swing drive unit and an air pressure control unit. The jet group is responsible for spraying airflow to the area where the block drop feature is located, the swing drive unit adjusts the angle of the jet nozzle, and the air pressure control unit adjusts the pressure of the airflow. Through the coordinated work of these components, the block drop can be effectively driven to move, thereby verifying whether the block drop feature actually exists.

[0075] The advantage of this system is that it can automatically detect and verify the characteristics of track block drop in the railway track area, reducing the workload and false detection rate of manual inspections. Compared with existing technologies, this system improves the accuracy and efficiency of detection through the combination of multiple sensors and algorithms, especially has stronger adaptability in complex environments. As a result, track block drop problems can be discovered and handled more promptly, ensuring the safety and reliability of railway operations.

Claims

1. A method for detecting falling blocks in a railway track area, characterized in that: The steps include: Track area image acquisition step, collecting the track area image of the train as the sleeper image through a visual camera; a step of detecting a falling block image, extracting falling block features in the sleeper image based on a preset image edge detection algorithm model, wherein the image edge detection algorithm model is configured to integrate Canny edge detection and multi-scale gradient feature extraction; a block drop verification step, when the block drop feature is detected, activating a block drop verification device, wherein the block drop verification device includes an adjustable air jet group for spraying air toward the area where the block drop feature is located to drive the block to move; a continuous image acquisition step, after the block drop verification device is started, continuously acquiring images of the track area as verification images; The step of comparing and outputting the position of the fallen block is to align the sleeper image and the verification image, divide the coordinate system based on the two tracks, and compare the relative position offset of the fallen block feature in the two track coordinate systems. If the offset exceeds a preset threshold, it is determined to be a real fallen block and the marking result is output; The steps for verifying the chip drop feature image are as follows: edge contours of the chip drop features in the sleeper image and the verification image are extracted respectively, and the contour similarity and area change rate of the two are calculated; if the contour similarity is lower than the preset value and the area change rate does not exceed the threshold, it is determined to be a static foreign object rather than a chip drop; if the contour similarity meets the requirements and the area change rate is associated with the position offset, it is determined to be a real chip drop.

2. The method for detecting falling blocks in a railway track area according to claim 1, characterized in that: The sleeper image and verification image are processed by a partitioning algorithm, and are divided into several parallel partitions with two tracks as the boundary. Each partition is parallel to the track direction and has an adjustable width. Each partition corresponds to an independent position code; When the block drop feature is detected to be located in a certain partition, the jet group of the block drop verification device is oriented and adjusted to match the spatial coordinates of the partition.

3. The method for detecting falling blocks in a railway track area according to claim 2, characterized in that: The block drop verification device includes at least two jet groups, which are respectively deployed on both sides of the track and symmetrically distributed on both sides of the track centerline; Each jet group includes an air nozzle, a swing component that drives the air nozzle's deflection angle, and a gas source for providing high-pressure gas to the air nozzle. The block drop verification device is equipped with an air path control system. When the block drop feature is located in a partition on one side of the track centerline, the jet group close to this side is started, and the swing component adjusts the jet angle of the jet nozzle according to the partition coordinates, and the jet nozzle drives the airflow to act on the block drop feature area.

4. The method for detecting falling blocks in a railway track area according to claim 3, characterized in that: The partition width of the verification image is strictly equal to the partition width of the sleeper image, and the partition boundaries are aligned by an image registration algorithm; In the step of comparing and outputting the position of the fallen block, if the offset of the fallen block feature in the verification image relative to the sleeper image spans at least one partition, it is determined to be a valid movement, and a fallen block motion trajectory report is generated by combining the air pressure parameters of the jet group and the partition offset distance.

5. The method for detecting falling blocks in a railway track area according to claim 4, characterized in that: It also includes the steps of storing and classifying lost blocks, specifically: Establish a block feature database to store the edge contour, motion trajectory, air pressure parameters and partition offset data of real block drops; Cluster analysis of block drop contour features is performed through convolutional neural networks to generate block drop type labels classified by shape and size, and then matched with historical data to optimize the judgment threshold.

6. The method for detecting falling blocks in a railway track area according to claim 5, characterized in that: It also includes the secondary verification step of dropped blocks, specifically: After the block loss verification device is started in the first block loss verification step, if the block loss feature of the verification image does not move, the secondary verification mode is triggered; In the secondary verification mode, the air path control system increases the air pressure of the jet group to a preset high pressure threshold, and adjusts the jet nozzle angle through the swing component to perform secondary injection; After the secondary injection, a secondary verification image is obtained through the continuous image acquisition step. If the block-drop feature still does not move, it is determined to be a static foreign body and the feature is marked as a non-block-drop type; A foreign object database is also established to store the location, air pressure parameters and image features of static foreign objects, and is updated synchronously to the railway operation and maintenance management system to eliminate false detection interference.

7. The method for detecting falling blocks in a railway track area according to claim 6, characterized in that: The block drop image detection step also includes calculating the pixel area, aspect ratio and geometric center coordinates of the block drop feature in the image, and performing spatial mapping based on the actual size of the track to output the estimated physical size of the block drop and the coordinates of the block drop location.

8. The method for detecting falling blocks in a railway track area according to claim 7, characterized in that: The block-dropping image detection step is configured with a dynamic air pressure adaptation algorithm, specifically including: In the block area quantification sub-step, the pixel area S of the block feature in the image is calculated by the edge detection processing module, and converted into a physical area S1 based on the mapping relationship between the actual track size and the image resolution; In the air pressure calculation sub-step, the basic air pressure threshold value P0 is set, and the first block drop verification injection pressure value P is dynamically adjusted according to the following formula: ， Among them, S2 is the preset reference area threshold, is the area weight coefficient, and its value range is 0.2 ≤ α ≤ 0.5; In the secondary verification mode, the secondary injection gas pressure is 1.2 to 2 times of p.

9. A railway track area block drop detection system, applying the railway track area block drop detection method according to any one of claims 1 to 8, characterized in that: include: Track image acquisition module, used to obtain sleeper images and verify images through visual cameras; An edge detection processing module configured to run an image edge detection algorithm model to extract block features; The jet verification execution module includes a symmetrically arranged jet group, a swing drive unit, and an air pressure control unit, which is used to directional drive the block to move; Image analysis and comparison module, used to register images, divide regions and calculate position offsets; The data storage and output module is used to store the block drop feature data and generate the detection report.

Images