Track material degradation diagnostic system

Abstract

To provide a track material deterioration diagnostic system that diagnoses a deterioration state of ballast without taking the trouble to visit the actual place by analyzing a track photographic image acquired by a monitoring device.SOLUTION: In a track material deterioration diagnostic system which analyzes a deterioration state of track material based upon a track photographic image, an analyzer device comprises: area division means which utilizes known position information on a rail fastening device on a track and a value of railroad tie width to divide the track photographic image into a railroad tie area and a ballast area; binarization means which performs binarization processing to use a first threshold to discriminate pixels of the railroad tie area between white and black and then performs binarization processing to use a second threshold to binarize pixels of the ballast area; and determination means which determines an image in which the mean luminance of the pixels of the ballast area is equal to or larger than a preset predetermined value as an image including defective ballast.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a system for diagnosing deterioration of track materials, and particularly to a technology effective for application to a system for diagnosing the deterioration state of ballast (crushed stones and gravel) in a ballast track using data acquired by a monitoring device mounted on a vehicle.

Background Art

[0002] In a ballast track, deterioration and defects of the ballast cause track displacement and train sway. Also, since it is greatly related to the occurrence of rail overhang accidents, appropriately grasping and managing defective portions of the ballast is a very important factor in track maintenance work. Currently, the management of ballast tracks is performed by visual inspection, for example, by inspecting the roadbed once a year or by comprehensively patrolling the track once every three months. In addition, a system that mounts a track equipment monitoring device under the vehicle floor to automatically determine damage and deterioration of track materials has been put into practical use on some lines, and it is possible to automatically determine the detachment of bolts of rail fastening devices and joint plates from the captured images. However, with regard to the crushed stones that make up the ballast track, a technique for determining the deterioration state from images has not been established.

[0003] Conventionally, as inventions related to devices and methods for detecting abnormalities in railway tracks and wear of track ballast, there are those described in Patent Documents 1 and 2, for example. Among these, the invention described in Patent Document 1 has an extraction means for extracting the edges in the longitudinal direction of the rail from an image of the railway track taken from above the rail, a calculation means for calculating the distribution in the longitudinal direction of the rail of the edge integration value obtained by integrating the edges in a direction substantially orthogonal to the longitudinal direction of the rail, and a recognition means for recognizing a region where the edge integration value is low in the distribution as a depression region. It recognizes the depression region from the track image and detects abnormalities in the railway track using the result.

[0004] On the other hand, the invention of the method for determining wear of track ballast described in Patent Document 2 comprises a striking process in which the ballast is struck, a vibration detection process in which vibrations propagated through the ballast by the striking process, and a deterioration analysis and determination process in which the frequency characteristics of the vibrations detected in the vibration detection process are analyzed and the degree of ballast wear is estimated by comparing the analyzed frequency characteristics with a reference value. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2006-176071 [Patent Document 2] Japanese Patent Publication No. 2015-117498 [Overview of the project] [Problems that the invention aims to solve]

[0006] The track image analysis device described in Patent Document 1 can recognize the sleeper area from images of the track and use the results to detect abnormalities in the railway track, but it has the problem that it cannot detect the deterioration state of the ballast. Furthermore, while the invention of the track ballast wear determination method described in Patent Document 2 allows for the determination of ballast wear status regardless of individual skill level, it includes a striking process that applies impact to the ballast and a vibration detection process. Therefore, determining the ballast wear status requires going to the site with an impact device and a vibration detection device to perform the work, which presents a challenge as it requires a great deal of effort.

[0007] This invention was made in view of the above-mentioned problems, and aims to provide a track material deterioration diagnosis system that can diagnose the deterioration state of ballast without having to go to the site, by analyzing track images acquired by a monitoring device equipped with an imaging device mounted on a vehicle. [Means for solving the problem]

[0008] To achieve the above objectives, this invention provides: A track material deterioration diagnostic system equipped with a display device and an analysis device having image processing functions, which analyzes the deterioration state of track materials based on track images acquired by a monitoring device mounted under the floor of a vehicle running on a track and linked to kilometer information, The aforementioned analysis device is A region division means that divides the track image into a sleeper region and a track bed region using known positional information of rail fastening devices on the track and the value of sleeper width, A binarization processing means that performs a binarization process to discriminate pixels in the sleeper region as white or black using a first threshold, and performs a binarization process on pixels in the track bed region using a second threshold, The system includes a determination means for determining that there is a defect in the track bed if the average brightness of the pixels in the track bed region is above a predetermined value set in advance.

[0009] With the above configuration, the pixels of the sleeper region and the ballast region, which are separated and extracted from the track image, are binarized using a first threshold and a second threshold, respectively. If the average brightness of the pixels in the ballast region is above a predetermined value, it is determined that there is a defect in the ballast shown in the image. By analyzing images captured by a camera mounted on a vehicle, the deterioration state of the ballast can be diagnosed without having to go to the site.

[0010] Preferably, the analysis device displays the image determined to be defective by the determination means and the image obtained by the binarization processing means in parallel on the screen of the display device. With this configuration, the image determined to be defective and the binarized image are displayed side-by-side on the display device screen, allowing the user to visually confirm whether the diagnosis is correct by comparing both images.

[0011] Furthermore, preferably, the analysis device is Object extraction means for extracting regions where white pixels are clustered from the binarized image as objects (targets for determination), An object area calculation means for calculating the area of ​​the range occupied by each of the aforementioned objects, The determination means includes the following: The determination means excludes images in which the total area calculated by the object area calculation means is equal to or greater than a predetermined value from the determination of whether or not there are defects in the track bed, either by determining that there are no defects in the track bed or by determining that it is impossible to determine. Perform exclusion process I will try to do so.

[0012] In railway tracks, there are areas where widespread whitening occurs due to mud ejection, or where the entire image is brightened by external lighting, making it impossible to identify the whitened crushed stone. With the above configuration, if the area determined to be white is larger than a predetermined value, other judgment processes can be omitted, thereby shortening the time required for analysis processing without reducing diagnostic accuracy.

[0013] Furthermore, preferably, the first threshold is the smaller of the maximum value of the threshold determined by Otsu's binarization method based on the brightness of the pixels in the sleeper region, or the brightness value determined based on the ratio of the number of pixels with the same brightness for each pixel in the sleeper region, and is the largest value among the values ​​obtained for multiple sleeper regions in the image. The second threshold is set to be the maximum value among the thresholds determined for each image by Otsu's binarization method based on the brightness of the pixels in the track bed region. Furthermore, preferably, the larger of the first threshold and the second threshold is used as the binarization threshold. This configuration makes it possible to improve the accuracy of diagnosing whether or not ballast deterioration has occurred.

[0014] Furthermore, preferably, the analysis device is An average luminance calculation means for calculating the average luminance of the pixel group before the binarization process corresponding to the range occupied by each of the objects and the average luminance of the pixel group outside the range occupied by each of the objects in the image before the binarization process, respectively; A most frequent value acquisition means for acquiring the value of the luminance most frequently represented in the image before the binarization process; A luminance difference calculation means for calculating the difference between the average luminance of the object range calculated by the average luminance calculation means and the most frequent value of the luminance obtained by the binarization process; It is provided with the above, and the determination means determines an image in which the value of the difference calculated by the luminance difference calculation means is equal to or greater than a preset predetermined value as having a defect in the ballast bed. According to such a configuration, it is possible to detect a track location where ballast deterioration has occurred with high accuracy while avoiding over-detection.

[0015] Furthermore, preferably, the determination means Objects included in the image that were not excluded from the determination of whether or not there is a track bed defect in the exclusion process described above, For a plurality of objects extracted from one image by the object extraction means, when there is an object larger than a preset area value among those whose areas have been calculated by the object area calculation means, it is determined that there is a defect in the ballast bed. According to such a configuration, the accuracy of diagnosing ballast deterioration can be improved.

[0016] Also, preferably, the first threshold value is a threshold value determined by Otsu's binarization method based on the luminance of the pixels in the sleeper area, or the number of pixels having the same luminance is obtained for each pixel in the sleeper area, and the smaller value of the luminance values that divide the top 10% and the remaining 90% is taken, and it is the largest value among the values obtained for a plurality of sleeper areas in the image. The analysis device includes a pattern recognition means for storing the range of the object in the image when the object extracted by the object extraction means is recognized as a character or symbol by performing determination by pattern matching using a sample image on the image binarized by the first threshold value. The determination means determines whether there is a defect in the roadbed in consideration of the recognition result by the pattern recognition means.

[0017] According to the above configuration, it is possible to determine whether characters or symbols are written on the surface of the sleeper, and by excluding the characters or symbols written on the surface of the sleeper from the determination target of whitening, the accuracy of diagnosing ballast deterioration can be improved.

[0018] Furthermore, preferably, the data acquired by the monitoring device includes 4m chord track displacement data associated with kilometer information. The analysis device On the XY orthogonal coordinates with the 4m chord track displacement on the horizontal axis and the area of the object on the vertical axis, a graph plotting dots indicating the area of each object in the image and the value of the 4m chord track displacement at the position of the object can be displayed on the screen of the display device. According to such a configuration, it is possible to determine whether there is a defect in the roadbed without requiring skill by looking at the displayed graph.

Effect of the Invention

[0019] According to the track material deterioration diagnosis system of the present invention, by analyzing the track captured image acquired by the monitoring device having an imaging device and mounted on a vehicle, it is possible to diagnose the deterioration state of the ballast without having to go to the site specifically.

Brief Description of the Drawings

[0020] [Figure 1] It is a system configuration diagram showing a configuration example of a track material deterioration diagnosis system according to an embodiment of the present invention. [Figure 2] It is a flowchart showing the first half of the deterioration diagnosis processing procedure in the track material deterioration diagnosis system according to an embodiment of the present invention. [Figure 3] It is a flowchart showing the second half of the deterioration diagnosis processing procedure of the embodiment. [Figure 4]The graph shows the number of occurrences of pixels in the entire image and the number of occurrences of pixels within the object area contained therein, with brightness on the horizontal axis and frequency on the vertical axis. (A) shows the case where the difference in average brightness between the object area and the surrounding area is large, and (B) shows the case where the difference in average brightness between the object area and the surrounding area is small. [Figure 5] This figure shows an example of the side-by-side display of a binarized image and the original image shown on a display device when a defect is detected in the crushed stone of the track bed. [Figure 6] This graph shows each object in an image represented by a dot, with the horizontal axis representing orbital displacement and the vertical axis representing object size. [Modes for carrying out the invention]

[0021] Hereinafter, with reference to the drawings, an embodiment of the present invention will be described in detail, in which the deterioration of track ballast (hereinafter referred to as "crushed stone"), which is composed of crushed stone, gravel, etc., is diagnosed using data acquired by a track equipment monitoring device mounted under the floor of a railway vehicle. Figure 1 shows an example configuration of one embodiment of the track material deterioration diagnosis system according to the present invention. As shown in Figure 1, the track material deterioration diagnosis system of this embodiment consists of a track equipment monitoring device 10 mounted on the underside of the vehicle, etc., and a server 20 and analysis device 30 on the ground side.

[0022] The track equipment monitoring device 10 consists of a track material monitoring device 11 equipped with a camera or the like as an imaging device to photograph the track surface from above, and a track displacement measurement device 12 equipped with an accelerometer or the like. The amount of rail displacement can be obtained by differentiating the data obtained by the accelerometer twice. The image data taken by the camera while the vehicle is running and the displacement data obtained by the accelerometer are linked to kilometer information indicating the vehicle's position at the time of data acquisition and stored in the storage unit 13 of the track equipment monitoring device 10.

[0023] The camera used to acquire images of the object to be analyzed by the diagnostic system of the present invention can be of any type as long as it is capable of capturing images that include the brightness values ​​of the object. However, in this embodiment, two line sensor cameras are used to output grayscale image data in which the brightness of the object's surface is represented by brightness values. One line sensor camera is set up so that it can capture an area of, for example, 2.5 m in the length direction of the rail and 70 cm in the width direction of the rail, and has a pixel count of, for example, 2.56 million pixels.

[0024] The analysis device 30 is implemented by installing a diagnostic program having the analysis algorithm described below onto a data processing device that has a configuration similar to a normal computer system, equipped with a processing unit such as a microprocessor (CPU), storage devices such as RAM, ROM, and a hard disk driver. The analysis device 30 is also connected to an input device 31 such as a keyboard or mouse, and a display device 32 such as an LCD panel.

[0025] Furthermore, the server 20 is equipped with a database 21 that stores information about the track. The database 21 stores information (in kilometers) about the location of the rail fastening devices (more precisely, the location of the bolts of the fastening devices) and information about the type of track bed (e.g., crushed stone, sieved gravel, concrete track bed, bridge, etc.). The track bed type information is also stored in the database 21, linked to the kilometers. In this embodiment, the location of the bolts of the rail fastening devices is the center in the width direction of the sleeper in the track bed being diagnosed. The data stored in the memory unit 13 of the track equipment monitoring device 10 is wirelessly transmitted to and stored in the ground-side server 20.

[0026] Next, an example of a specific procedure for diagnosing the deterioration of ballast in the track material deterioration diagnosis system according to the present invention will be explained using the flowcharts in Figures 2 and 3. When starting the process according to this flowchart, the analysis device 30 requests the server 20 to transmit track images and rail displacement data to be diagnosed via a communication network such as the Internet, and reads the data.

[0027] In the deterioration diagnosis method for track bed crushed stone of this embodiment, as shown in Figure 2, first, the kilometer information associated with the track photograph image to be read is used as an indicator to refer to the database 21 of the server 20 and determine the type of track bed (e.g., crushed stone, sieved gravel, concrete track bed such as level crossings, bridges, etc.) (Determination 1; Step S1). If the track bed is anything other than crushed stone, it is determined that there is no defect. When the determination of the type of track bed is completed for one image, the determination of the type of track bed is repeated for the next image (Process 1). In Step S1, if the track bed is anything other than crushed stone, it may be excluded from the following processes as diagnosis is not possible.

[0028] Once the process of determining the track bed type is completed for all images to be diagnosed, the process proceeds to the next step. Using the kilometer information associated with the image, the database is referenced to obtain the position information of the rail fastening devices (bolts) contained in the image (Process 2; Step S2). If obtaining the position information fails, it is determined that there is a defect, and the diagnostic process for the next image is initiated. Next, the image is divided into a sleeper region and a ballast region by identifying the sleeper region within the image using the position information (center of the sleeper) obtained in step S2 and the pre-entered sleeper width, and considering the rest as the ballast region (processing 3; step S3). If the region division fails at this stage, it is determined that there is a defect, and the process moves on to diagnosing the next image.

[0029] On the other hand, if the region division is successful in step S3, the process proceeds to the next step, for example, by calculating a threshold for binarizing pixels in the sleeper region using Otsu's binarization method (process 4; step S4). Also, the number of pixels with the same brightness is determined for each pixel in the sleeper region and the threshold for binarization is set based on their ratio (distribution) (process 5; step S5). Next, the smaller of the threshold calculated in process 4 and the threshold determined in process 5 is recorded (process 6; step S6). Then, processes 4 to 6 are performed for all sleeper regions in the image. Since Otsu's binarization method is a well-known technique for automatically determining thresholds in image processing, a detailed explanation will be omitted.

[0030] Next, the entire image is binarized using the threshold recorded in step S6 (process 7; step S7). Subsequently, character recognition is performed on the white areas of the railway tie region in the binarized image using pattern matching with template images of various characters (including numbers and symbols) (process 8; step S8), and the location information (coordinate information in the image) of the recognized characters is recorded (process 9; step S9). Then, processes 7 to 9 are performed for all images to be diagnosed. Steps S4 to S9 are preprocessing steps for the railway tie region in the image.

[0031] Once the above preprocessing is complete, the process proceeds to the next step, where a threshold for pixel binarization is calculated for the track bed region within the image divided in step S3 using Otsu's binarization method (processing 10; step S10). Next, the average brightness of the track bed region is calculated (processing 11; step S11), and the calculated average brightness is compared with the pre-inputted and set average value of the track bed region brightness to determine whether it is above or below the average value (determination 2; step S12). If the average brightness is above the average value, it is determined that there is a defect in the crushed stone of the track bed in that image. Then, steps S10 to S12 are repeated for all images (processing 12).

[0032] Subsequently, for each image that was not determined to have a defect in step S12, the maximum value among the Otsu thresholds calculated in process 10 for each track bed region and the maximum value among the smaller thresholds recorded in process 6 for each sleeper region are extracted (process 13; step S13). Next, the larger of the two maximum values ​​extracted in step S13 is adopted and recorded as the value for binarization of the entire image region (process 14; step S14). By determining the threshold for binarization in this way, the accuracy of the diagnosis can be improved.

[0033] Next, each image is binarized using the values ​​adopted in step S14 above, and the parts (collections) separated into white are extracted as objects (objects for analysis) (processing 15; step S15). Subsequently, it is determined whether the total area of ​​the objects in the binarized image is greater than or less than a predetermined maximum value that has been input and set in advance (determination 4; step S16). If the total area of ​​the objects is determined to be larger than the set value, the crushed stone of the track bed in that image is deemed to have no defects. In railway tracks, there are areas where widespread whitening occurs due to mud spray, or the entire image is brightened by external lighting, making it impossible to identify the whitened crushed stone. Therefore, such images are excluded from the analysis. In step S16, instead of determining that there are no defects, the image may be excluded from the determination of whether or not there are defects by marking it as "undetermined."

[0034] Next, the average brightness of pixels within the corresponding area of ​​the original image corresponding to the range of the object extracted in step S15 is calculated (processing 16; step S17). Subsequently, the brightness value that appears most frequently throughout the entire image is calculated as the mode (processing 17; step S18). Then, it is determined whether the difference between the average brightness of the object range calculated in step S17 and the mode calculated in step S18 is greater or less than the pre-inputted and set average brightness difference between the object range and its surroundings (determination 5; step S19). If it is determined that the difference between the average brightness of the object range and its mode is less than the set value for the average brightness difference between the object range and its surroundings, then the crushed stone of the track bed in that image is deemed to have no defects.

[0035] Figures 4(A) and 4(B) show the number of occurrences of the brightness of pixels in the entire image and the brightness of pixels within the object area, with brightness on the x-axis and frequency on the y-axis, for cases where the difference in average brightness between the object area and its surroundings is large and for cases where the difference in average brightness between the object area and its surroundings is small. Figure 4(A) shows the case where the difference in average brightness between the object area and its surroundings is large, while Figure 4(B) shows the case where the difference in average brightness between the object area and its surroundings is small, i.e., when the image change due to whitening is difficult to discern.

[0036] When the difference in average brightness between the object area and the surrounding area is large, as shown in Figure 4(A), the curve a of the entire image and the curve b of the object area have significantly different patterns, and the difference between the mode of brightness in the entire image and the average brightness of the object area becomes large. On the other hand, when the difference in average brightness between the object area and the surrounding area is small, as shown in Figure 4(B), the curve a' of the entire image and the curve b' of the object area have similar patterns, and the difference between the mode of brightness in the entire image and the average brightness of the object area becomes small. The judgment 5 in step S19 uses these characteristics to judge the image, and can avoid over-detection of whitened areas.

[0037] Next, the size (area) of each object in the image extracted in step S15 is calculated, and it is determined whether the calculated area is larger or smaller than the pre-entered and set object area value (Determination 6; step S20). If it is determined that there are no objects larger than the set value, it is determined that there are no defects in the crushed stone of the track bed in that image. This is because it is known that in images of tracks where the crushed stone is whitening, the size of the objects will be above a certain size, and if the objects are below a certain size, it is thought to be due to noise or the like.

[0038] Next, for images that were determined in step S20 to contain objects of a certain size or larger, it is determined whether the area where the characters recorded in step S9 were recognized overlaps with the area of ​​the object detected in step S20 (determination 7; step S21). If it is determined that the area where the characters were recognized and the area of ​​the object overlap, it is determined that there are no defects in the crushed stone of the track bed in that image. This is because the objects extracted from the image are thought to be characters and symbols written on the surface of the sleepers. On the other hand, if it is determined that the area where the characters were recognized and the area of ​​the object do not overlap, it is determined that there are defects in the crushed stone of the track bed in that image, and the determination result is recorded along with the kilometer information of the image capture location. Then, steps S15 to S21 are repeatedly performed for all images (processing 18).

[0039] In the diagnostic system of this embodiment, by performing processing according to the procedure described above, it is possible to detect track sections where whitening of the crushed stone is occurring with high accuracy while avoiding false positives. Furthermore, in the diagnostic system of this embodiment, for images in which defects are determined to be present in the crushed stone of the track bed, a list of kilometer information indicating the track section where the defect was determined is displayed on the display device screen. When a kilometer is specified on that list, the binarized image and the original image are displayed side by side, as shown in Figure 5. In this way, the original image in which the defect was determined and the binarized image are displayed side by side on the display device screen, allowing the user to visually confirm whether the diagnostic result is correct by comparing both images.

[0040] Furthermore, based on their many years of experience, the inventors hypothesized that poor crushed stone quality is closely related to track sway (dynamic settlement due to train wheel load). They investigated the correlation between various displacement data obtained by a track displacement measurement device, which constitutes a track equipment monitoring system, and track whitening, and found that 4m chord track displacement data showed the highest correlation with track whitening.

[0041] The analysis device of this embodiment uses kilometer information indicating track locations where defects in the crushed stone of the track bed have been determined to be present, to read 4m chord track displacement data measured by a track displacement measurement device from a database. Then, based on this 4m chord track displacement data and the area (size) of the object calculated in step S20, the device is configured to create a graph like the one shown in Figure 6, with track displacement on the horizontal axis and object size on the vertical axis, representing each object as a dot, and display this graph on the screen of the display device, through a program.

[0042] By looking at the graph shown in Figure 6, defects in the track bed can be detected without requiring any special skills. Furthermore, the program is configured so that when you use the mouse to select any dot on the graph shown in Figure 6 on the display screen and click the mouse button, a list screen displaying kilometer information and image numbers will appear. Selecting an item from the list will then display detailed information about the selected location.

[0043] Although the present inventors' invention has been described in detail above based on embodiments, the present invention is not limited to the above embodiments. For example, in the above embodiments, the whitening of crushed stone as track material is diagnosed based on grayscale image data acquired by a track material monitoring device already installed on the train, but the deterioration of crushed stone may also be diagnosed by performing the same processing as described in the embodiments on images taken with a color camera.

[0044] Furthermore, in the above embodiment, a common threshold is set for multiple sleeper regions captured in a single image to perform binarization. However, since multiple sleepers are usually captured in a single image, a binarization threshold may be set for each sleeper region. In addition, information correlated with image whitening, such as 4m chord track displacement, black spots on the rail head, joint information, welding, and sleeper type, may be added to the judgment, and a score may be assigned from the combination of these to be incorporated into the judgment logic to improve the judgment accuracy.

[0045] Furthermore, since the condition of the rail head also correlates with the whitening of the image, it would be beneficial to perform judgment on the image of the rail head as well. Additionally, a function to automatically learn and store characters, symbols, and markings inscribed on the sleeper surface may be added. Furthermore, in the above embodiment, several settings are manually set according to the track conditions, but a function to mechanically set thresholds from defined defective images may also be added. In addition, a function to estimate the shape of the crushed stone from the brightness difference between the object and its surroundings and to determine the degree of wear of the crushed stone may also be added. Furthermore, a function to accumulate diagnostic results and analyze the sequential rate of deterioration may also be added. [Explanation of Symbols]

[0046] 10. Track Equipment Monitoring Device 11. Orbital material monitoring device 12. Track displacement detection device 13 Storage section 20 servers 21 Databases 30 Analyzer 31 Input device 32 Display device

Claims

1. A track material deterioration diagnostic system comprising a display device and an analysis device having image processing functions, which analyzes the deterioration state of track materials based on track images acquired by a monitoring device mounted under the floor of a vehicle running on a track and linked to mileage information, The aforementioned analysis device is A region division means that divides the track image into a sleeper region and a track bed region using known positional information of rail fastening devices on the track and the value of sleeper width, A binarization processing means that performs a binarization process to discriminate pixels in the sleeper region as white or black using a first threshold, and performs a binarization process to discriminate pixels in the track bed region using a second threshold, A determination means for determining that there is a defect in the track bed if the average brightness of the pixels in the track bed region is above a predetermined value set in advance, A track material degradation diagnostic system equipped with the following features.

2. The track material deterioration diagnosis system according to claim 1, wherein the analysis device displays in parallel on the screen of the display device an image determined to be defective by the determination means and an image obtained by binarizing the said image by the binarization processing means.

3. The aforementioned analysis device is Object extraction means for extracting regions where white pixels are clustered together as objects from the binarized image, An object area calculation means for calculating the area of ​​the range occupied by each of the aforementioned objects, The track material deterioration diagnosis system according to claim 2, comprising, wherein the determination means performs an exclusion process to exclude images in which the total value of the area calculated by the object area calculation means is greater than or equal to a predetermined value set in advance, from the determination of whether or not there are defects in the track bed, as determined to be either free from defects or undeterminable.

4. The first threshold is the smaller of the maximum value of the threshold determined by Otsu's binarization method based on the brightness of the pixels in the sleeper region, or the brightness value determined based on the ratio of the number of pixels with the same brightness for each pixel in the sleeper region, and is the largest value among the values ​​obtained for multiple sleeper regions in the image. The track material deterioration diagnostic system according to claim 3, wherein the second threshold is the maximum value among the thresholds determined by Otsu's binarization method based on the brightness of pixels in the track bed region for each of the multiple images.

5. The track material degradation diagnostic system according to claim 4, characterized in that the larger of the first threshold and the second threshold is used as the binarization threshold.

6. The aforementioned analysis device is An average brightness calculation means that calculates the average brightness of the pixel group before binarization corresponding to the range occupied by each of the objects, and the average brightness of the pixel group other than the range occupied by each of the objects in the image before binarization, A mode acquisition means for obtaining the brightness value that appears most frequently in the image before the aforementioned binarization process, A luminance difference calculation means calculates the difference between the average luminance of the object range calculated by the average luminance calculation means and the mode of luminance obtained by the binarization process, The track material deterioration diagnosis system according to claim 5, comprising, wherein the determination means determines that there is a defect in the track bed if the difference value calculated by the brightness difference calculation means is greater than or equal to a predetermined value set in advance.

7. The determination means is The track material deterioration diagnosis system according to claim 6, wherein, among the objects included in the image that were not excluded from the determination of whether or not there is a track bed defect in the exclusion process, if, among the multiple objects extracted from one image by the object extraction means, there is an object whose area is larger than a preset area value among those whose area is calculated by the object area calculation means, it is determined that there is a defect in the track bed.

8. The first threshold is the smaller of the threshold determined by Otsu's binarization method based on the brightness of the pixels in the sleeper region, or the brightness value determined based on the ratio of the number of pixels with the same brightness for each pixel in the sleeper region, and is the largest value among the values ​​obtained for multiple sleeper regions in the image. The analysis device includes a pattern recognition means that, when it performs a pattern matching determination using a sample image on an object extracted by the object extraction means from the binarized image using the first threshold, and recognizes the object as a character or symbol, stores the range of the object within the image. The track material deterioration diagnosis system according to claim 7, wherein the determination means determines whether or not there is a defect in the track bed, taking into consideration the recognition result by the pattern recognition means.

9. The data acquired by the monitoring device includes 4m chord trajectory displacement data linked to kilometer information. The aforementioned analysis device is A track material degradation diagnosis system according to any one of claims 1 to 8, wherein a graph is displayed on the screen of the display device, which plots the area of ​​each object in the image and a dot indicating the value of the 4m chord track displacement at the position of that object on an XY Cartesian coordinate system with the horizontal axis representing the 4m chord track displacement and the vertical axis representing the area of ​​the object.

Images