Trolley wire inspection device and trolley wire inspection method

Abstract

To provide a trolley wire inspection device and a trolley wire inspection method that accurately detect worn areas of a trolley wire.SOLUTION: A line sensor image generating unit 30 generates a line sensor image from the image captured by a line sensor 22 . A worn area extraction image generation unit 32 extracts a worn area of a trolley wire by performing image processing according to multiple image processing execution procedures M registered in an image processing execution procedure memory unit 33, corrects the worn area by scanning outward from the extracted worn area and by assigning a position with a brightness value below a predetermined brightness value as a boundary of the worn area, and generates multiple worn area extraction images corresponding to each of the multiple image processing execution procedures M. A worn area extraction image evaluation unit 34 selects the worn area extraction image with the highest evaluation value from among the multiple worn area extraction images. A trolley wire deviation calculation unit 35 and a trolley wire remaining diameter calculation unit 36 calculate a deviation and remaining diameter of the trolley wire based on the selected worn area extraction image.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to an overhead line inspection device and an overhead line inspection method.

Background Art

[0002] When an electric railway vehicle runs on a track, it obtains power by being supplied with electricity from an overhead line provided above the track via a current collector such as a pantograph. Every time the electric railway vehicle passes, the current collector slides on the lower surface of the overhead line. Therefore, if the electric railway vehicle is continuously operated, the overhead line gradually wears out and eventually breaks. Therefore, a wear limit is set for the overhead line, and when the remaining diameter of the overhead line falls below the management value, the overhead line is replaced with a new one. On the other hand, since the contact portion of the current collector also wears, the overhead line is installed in a zigzag manner in the direction perpendicular to the rail so that the contact portion does not concentrate in one place, and this displacement amount is also managed. It takes a great deal of labor for workers to manually inspect and measure these overhead lines.

[0003] For example, Patent Document 1 discloses an automatic inspection device including illumination that irradiates a worn area located below an overhead line above an electric railway vehicle, and a line sensor that images the worn area (sliding surface with which the pantograph comes into contact) of the overhead line irradiated by the illumination. More specifically, the line sensor is provided so that the scanning line crosses the overhead line upward, and an image captured by the line sensor is arranged in time series to generate a line sensor image. Then, after applying image processing to this line sensor image, for example, by using known data regarding the specifications of the overhead line such as the height and thickness of the overhead line, the displacement and remaining diameter of the overhead line are calculated.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in the above-described apparatus, due to lighting, the state of the worn area of the trolley wire, the imaging environment, etc., there may be a difference in the luminance of the worn area of the trolley wire in the line sensor image, and there is a problem that the worn area cannot be accurately acquired.

[0006] Therefore, an object of the present invention is to provide a trolley wire detection apparatus and a trolley wire detection method capable of accurately detecting a worn area of a trolley wire.

Means for Solving the Problems

[0007] In order to solve the above problems, the present invention employs the following means. That is, a trolley wire detection apparatus according to a first invention includes a line sensor that images a trolley wire, and based on an image captured by the line sensor as an input image, the trolley wire is detected. The trolley wire detection apparatus includes a line sensor image generation unit that generates a line sensor image by arranging the input images in time series, an image processing execution procedure storage unit in which a plurality of execution orders are registered as image processing execution procedures when sequentially executing a plurality of image processing functions on the line sensor image, and according to each of the plurality of image processing execution procedures, by executing the plurality of image processing functions with the line sensor image as an input, extracting a worn area of the trolley wire, scanning outward from the extracted worn area of the trolley wire to set a position where the luminance value becomes equal to or less than a preset luminance value as a boundary of the worn area to correct the worn area, and generating a plurality of worn area extraction images corresponding to each of the plurality of image processing execution procedures, a worn area extraction image evaluation unit that calculates an evaluation value for each of the plurality of worn area extraction images and selects the worn area extraction image with the highest evaluation value, and a trolley wire detection unit that calculates either or both of the remaining diameter and deviation of the trolley wire based on the selected worn area extraction image with the highest evaluation value.

[0008] In addition, in the trolley wire detection device according to the second invention, for each of the plurality of image processing execution procedures, 0 or more contrast adjustment processes, 0 or more background / noise removal processes, 1 or more binarization processes, and 0 or more interpolation processes are registered so as to be executed in this order.

[0009] In addition, in the trolley wire detection device according to the third invention, a priority is associated with and stored for each of the plurality of image processing execution procedures. When the same evaluation value is calculated for a plurality of the wear area extraction images, the wear area extraction image corresponding to the image processing execution procedure with the higher priority is selected as the wear area extraction image with the highest evaluation value.

[0010] In addition, in the trolley wire detection device according to the fourth invention, the wear area extraction image evaluation unit calculates the number of wear areas for each of the plurality of wear area extraction images, and calculates the evaluation value so that the evaluation value becomes higher when the difference between the number of wear areas and the actual number of trolley wires is small.

[0011] In addition, in the trolley wire detection device according to the fifth invention, the wear area extraction image evaluation unit calculates a trolley wire likelihood, which is an index indicating that the wear area has characteristics on an image close to the trolley wire, for each of the plurality of wear area extraction images, and calculates the evaluation value so that the evaluation value becomes higher when the trolley wire likelihood is high.

[0012] In addition, in the trolley wire detection device according to the sixth invention, the image processing execution procedure is registered in the image processing execution procedure storage unit corresponding to each of a plurality of different environments when the image is captured by the line sensor. The trolley wire detection device further includes an environment determination unit that determines the environment in which the input image was captured based on the line sensor image, and the wear area extraction image generation unit generates the wear area extraction image according to the image processing execution procedure corresponding to the determined environment.

[0013] Moreover, the trolley wire detection method according to the seventh invention is a trolley wire detection method for detecting the trolley wire based on an image captured by a line sensor that captures the trolley wire as an input image, including: arranging the input images in time series to generate a line sensor image; when sequentially executing a plurality of image processing functions on the line sensor image, the execution order is registered in plural as an image processing execution procedure, and according to each of the plural image processing execution procedures, executing the plural image processing functions with the line sensor image as an input to extract a worn area of the trolley wire; scanning outward from the extracted worn area of the trolley wire to set a position where the luminance value becomes equal to or less than a preset luminance value as a boundary of the worn area to correct the worn area; generating a plurality of worn area extraction images corresponding to each of the plural image processing execution procedures; calculating an evaluation value for each of the plural worn area extraction images and selecting the worn area extraction image with the highest evaluation value; and calculating either or both of the remaining diameter and deviation of the trolley wire based on the selected worn area extraction image with the highest evaluation value.

Advantages of the Invention

[0014] According to this invention, even when a luminance difference occurs in the worn area, the worn area of the trolley wire can be accurately detected.

Brief Description of the Drawings

[0015]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Embodiments for Carrying Out the Invention

[0016] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. FIG. 1 is an explanatory diagram showing a traveling environment of an inspection vehicle 1 for inspecting a trolley wire on which a trolley wire detection device 20 according to this embodiment is mounted. In FIG. 1, the inspection vehicle 1 travels on a pair of tracks 2 laid on the ground. A plurality of utility poles 3 are erected at intervals along the side of the track 2. The utility poles 3 are provided with beams 5 and curved fittings 6. Above the beam 5, a catenary wire 4, which is a power line for supplying power to the inspection vehicle 1, extends along the extending direction R of the track 2 and is supported by the beam 5. Below the beam 5 and the curved fittings 6, a hanging wire 7 for hanging the trolley wire 9 to be described next is supported by the beam 5 and the curved fittings 6 so as to extend along the extending direction R of the track 2. A plurality of hangers 8 are hung on the hanging wire 7 at substantially constant intervals in the extending direction R of the track 2.

[0017] The trolley wire 9 is hung on these plurality of hangers 8 so as to have a constant tension, and is provided below the hanging wire 7 and substantially parallel to the hanging wire 7. Electric power is supplied to the trolley wire 9 from the catenary wire 4 by a catenary branch wire (not shown). A current collection device 1b such as a pantograph is provided on the upper surface 1a of the inspection vehicle 1. When the current collection device 1b contacts the trolley wire 9 during the running of the inspection vehicle 1, electric power is supplied from the trolley wire 9 to the inspection vehicle 1.

[0018] The trolley wire 9 is provided such that, for example, one or two run parallel to a pair of tracks 2. In order to prevent a section without power supply from occurring when the trolley wire 9, which is the power supply source, switches to the next adjacent trolley wire 9 in the extending direction R of the track 2 during the running of an electric railway vehicle including the inspection vehicle 1, the trolley wire 9 is provided at its end to overlap with another adjacent trolley wire 9 along the extending direction R of the track 2. For this reason, in the present embodiment, there may be a section where up to four trolley wires 9 overlap and are provided with respect to a pair of tracks 2. In addition, when photographing a branching point of the trolley wire 9 corresponding to a branch of the track, a larger number of trolley wires 9 than in a normal section may be imaged.

[0019] When the trolley wire 9 is provided parallel to the track 2, the portion of the current collection device 1b that contacts the trolley wire 9 is limited, so there is a risk that the limited portion of the current collection device 1b will wear and be damaged unevenly. To prevent this, the trolley wire 9 is provided to meander gently in a zigzag along the extending direction R of the track 2. In the trolley wire 9, the displacement, which is the left-right deviation of the trolley wire 9 with respect to the center between the pair of tracks 2 caused by this meandering, is strictly controlled.

[0020] Each time an electric railway vehicle including the inspection vehicle 1 passes, the current collection device 1b slides on the lower surface of the trolley wire 9. For this reason, if the electric railway vehicle is continuously operated, the trolley wire 9 gradually wears and finally breaks. Therefore, a wear limit is set for the trolley wire 9, and when the remaining diameter of the trolley wire 9 falls below the control value, the trolley wire 9 is replaced with a new one.

[0021] In the trolley wire inspection device 20 according to this embodiment, the trolley wire 9 is inspected by measuring the remaining diameter and displacement of the trolley wire 9. The trolley wire inspection device 20 includes an illumination 21, a line sensor 22, and a control device 23. The illumination 21 is provided on the upper surface 1a of the inspection vehicle 1 so as to irradiate light upward. When the trolley wire 9 is imaged by the line sensor 22 described below, the wear region formed in a planar shape due to the wear of the trolley wire 9 caused by the sliding of the current collector 1b on the lower side of the trolley wire 9 reflects the light of the illumination 21, and thus is imaged so as to be displayed as a bright region.

[0022] The line sensor 22 is provided on the upper surface 1a of the inspection vehicle 1 in the same manner as the illumination 21. The line sensor 22 is a camera that images an image composed of one pixel in the vertical direction and a plurality of pixels, for example, 8192 pixels in the horizontal direction. In this embodiment, since the line sensor 22 images a grayscale image, each pixel has one luminance value.

[0023] The line sensor 22 is provided vertically upward and in such a manner that the direction 22b of the scanning line 22a is orthogonal to the extending direction R of the track 2 and is the same as the extending direction of the sleepers (not shown). Thereby, the scanning line 22a of the line sensor 22 crosses the trolley wire 9. The line sensor 22 continuously images the trolley wire 9 at short time intervals while the inspection vehicle 1 is running. The image captured by the line sensor 22 is transmitted to the control device 23 as an input image. The control device 23 is an information processing device such as a personal computer, for example. The control device 23 is mounted on the inspection vehicle 1.

[0024] FIG. 2 is a block diagram showing the functional configuration of the trolley wire inspection device 20 according to this embodiment. In FIG. 2, the control device 23 includes a line sensor image generation unit 30, an environment determination unit 31, a wear region extraction image generation unit 32, an image processing execution procedure storage unit 33, a wear region extraction image evaluation unit 34, a trolley wire displacement calculation unit (trolley wire inspection unit) 35, and a trolley wire remaining diameter calculation unit (trolley wire inspection unit) 36.

[0025] The line sensor image generation unit 30 receives an input image from the line sensor 22, arranges, for example, 1000 lines of this image in a time series in the vertical direction, and generates a line sensor image.

[0026] FIG. 3 is a schematic diagram showing an example of a line sensor image generated based on an image captured by the line sensor of the trolley wire detection device according to the present embodiment. In the line sensor image 50, the horizontal direction X is the direction 22b of the scanning line 22a of the line sensor 22, and the vertical direction Y is the time series direction, that is, the direction in which the input images are arranged.

[0027] In the line sensor image 50 shown in FIG. 3, the trolley wire 9 is imaged closest to the viewer. The trolley wire 9 has a substantially circular cross section in a new state, and its lower side is worn away by the sliding of the current collector 1b, resulting in a planar wear region 9a. Since the wear region 9a is irradiated with light by the illumination 21, it is imaged brightly. On both sides of the wear region 9a of the trolley wire 9, the surface 9b of the trolley wire 9 that curves upward while curving outward from the widthwise end of the wear region 9a is imaged. In the line sensor image 50 of FIG. 3, the portion where the trolley wire 9 is bent by the bending bracket 6 is imaged. Also, due to factors such as illumination, the state of the wear region of the trolley wire, and the imaging environment, a luminance difference may occur in the wear region 9a, and dark regions 9c, 9d with low luminance may occur partially.

[0028] Behind the trolley wire 9, the suspension wire 7 is imaged. Further behind the suspension wire 7, the contact wire 4 and the beam 5 and bending bracket 6 that support them are imaged. The beam 5 and the bending bracket 6 are imaged thinner than they actually are due to the influence of the imaging interval of the line sensor 22. The sky 10 is imaged as the background of the trolley wire 9, the suspension wire 7, the contact wire 4, the beam 5, and the bending bracket 6. In this line sensor image 50, since it is imaged in bright daylight, the sky 10 is displayed brightly with a high luminance value.

[0029] The line sensor image generation unit 30 generates such a line sensor image 50 and transmits it to the wear area extraction image generation unit 32.

[0030] Based on the line sensor image 50 generated by the line sensor image generation unit 30, the environment determination unit 31 determines the environment in which the input image was captured. The environment determination unit 31 determines five types of environments: night or tunnel, dark daytime, bright daytime, the moment of entering a tunnel during the day, and other situations. For this determination, the environment determination unit 31 calculates the histogram of the line sensor image 50 and calculates the average value, the most frequent value, the second highest luminance value, and the third highest luminance value. Then, it compares each of the calculated values with each of the previously set night determination threshold, dark daytime determination threshold, and bright daytime determination threshold as follows to determine the environment.

[0031] When all of the following determination expressions are satisfied simultaneously, the environment determination unit 31 determines that the environment corresponding to the line sensor image 50 is night or tunnel. 0 ≦ average value ≦ night determination threshold 0 ≦ most frequent value ≦ night determination threshold 0 ≦ second highest luminance value ≦ night determination threshold 0 ≦ third highest luminance value ≦ night determination threshold

[0032] When all of the following determination expressions are satisfied simultaneously, the environment determination unit 31 determines that the environment corresponding to the line sensor image 50 is dark daytime. night determination threshold < average value ≦ dark daytime determination threshold night determination threshold < most frequent value ≦ dark daytime determination threshold night determination threshold < second highest luminance value ≦ dark daytime determination threshold night determination threshold < third highest luminance value ≦ dark daytime determination threshold

[0033] When all of the following determination expressions are satisfied simultaneously, the environment determination unit 31 determines that the environment corresponding to the line sensor image 50 is bright daytime. bright daytime determination threshold < average value Bright daytime determination threshold < mode value Bright daytime determination threshold < second most frequent luminance value Bright daytime determination threshold < third most frequent luminance value

[0034] When all of the following determination expressions are simultaneously satisfied, the environment determination unit 31 determines that the environment corresponding to the line sensor image 50 is the moment of entering a tunnel during daytime. Night determination threshold < average value ≤ maximum value of luminance values (for example, 255) 0 ≤ mode value ≤ night determination threshold 0 ≤ second most frequent luminance value ≤ night determination threshold 0 ≤ third most frequent luminance value ≤ night determination threshold

[0035] When not corresponding to any of the above environments, the environment determination unit 31 determines that it is another environment. The environment determination unit 31 transmits the determined environment result to the wear area extraction image generation unit 32.

[0036] The wear area extraction image generation unit 32 receives the line sensor image 50 from the line sensor image generation unit 30 and the determination result (environment) from the environment determination unit 31, and extracts the wear area 9a of the trolley wire 9 by executing various image processing functions on the line sensor image 50, thereby generating a wear area extraction image.

[0037] The wear area extraction image generation unit 32 includes an image processing unit 40 and a post-processing unit 41. The image processing unit 40 is configured such that various image processing functions can be selectively executed. The image processing unit 40 includes a contrast adjustment processing unit 40A, a background / noise removal processing unit 40B, a binarization processing unit 40C, an interpolation processing unit 40D, and a wear area correction unit 40E.

[0038] The contrast adjustment processing unit 40A is configured such that one or more image processing functions for the purpose of adjusting the gradation of an image can be selectively executed. As the contrast adjustment processing, for example, gamma correction processing for correcting the gradation of an image to an appropriate curve according to the gamma value, or processing for adjusting the gradation by normalizing the luminance value can be applied. Note that the contrast adjustment processing included in the contrast adjustment processing unit 40A is not limited to that shown above.

[0039] The background / noise removal processing unit 40B is configured such that one or more image processing functions for the purpose of removing an object different from the overhead line, such as a building imaged in the background, or noise can be selectively executed. As the background / noise removal processing, for example, GSTH (gray-scale top-hat) processing can be applied in which a gray-scale image is shrunk, dilated, opened, and then the original image before dilation and contraction is subtracted. Note that the background / noise removal processing included in the background / noise removal processing unit 40B is not limited to that shown above.

[0040] The binarization processing unit 40C is configured such that one or more image processing functions for the purpose of converting an image with a plurality of gradations into, for example, two gradations of white and black and extracting only the worn area 9a can be selectively executed. As the binarization processing, for example, discriminant analysis binarization processing can be applied in which a histogram of luminance values is generated for the image, and the binarization threshold is determined according to the image so that the variance ratio of the within-class variance and between-class variance for each of the background and the worn area 9a is maximized. Note that the binarization processing included in the binarization processing unit 40C is not limited to that shown above.

[0041] The interpolation processing unit 40D is configured such that one or more image processing functions for the purpose of interpolating the shape of the worn area 9a by joining together areas that are supposed to be the original single worn area 9a but are divided into a plurality of areas when an image in which the binarization processing in the binarization processing unit 40C is executed and an area assumed to be the worn area 9a is extracted is targeted, can be selectively executed.

[0042] As the interpolation process, for closing process that joins regions that are segmented and represented adjacent to each other in the horizontal direction X by expanding and contracting the image in the horizontal direction X, and for wavy wear interpolation process that joins regions when the regions are segmented and represented in a striped or wavy pattern in the vertical direction Y, these can be applied. Note that the interpolation process included in the interpolation processing unit 40D is not limited to those shown above.

[0043] FIG. 4 is a schematic diagram showing an example of a wear region extraction image temporarily acquired from the trolley wire detection device according to the present embodiment, and as described above, it is the wear region extraction image 50A processed by the interpolation processing unit 40D. In the wear region extraction image 50A, due to lighting, the state of the wear region 9a of the trolley wire 9, the shooting environment, etc., as shown in FIG. 3, a luminance difference occurs in the wear region 9a, and there may be dark regions 9c, 9d where the luminance is lower than other regions. That is, in the wear region extraction image 50A processed by the interpolation processing unit 40D, the dark regions 9c, 9d are not detected as the wear region 9a, so the accurate wear region 9a cannot be detected.

[0044] As shown in FIG. 4, in the wear region extraction image 50A processed by the interpolation processing unit 40D, only the brightly imaged wear region 9a is extracted. Therefore, in the present embodiment, the wear region extraction image 50A processed by the interpolation processing unit 40D is corrected so as to detect the dark regions 9c, 9d as the wear region 9a. The wear region correction unit 40E performs a raster scan outward from the position of the wear region extracted by the extraction process of the wear region 9a for each line in the vertical direction of the image, and sets the position where the luminance value becomes equal to or lower than a preset luminance value as the boundary between the sliding surface and the non-sliding surface, that is, the boundary of the wear region 9a.

[0045] FIG. 5 is a schematic diagram showing an example of a wear area extraction image finally obtained by the trolley wire detection device according to the present embodiment. As a result of the above-described correction, since the dark areas 9c and 9d shown in FIG. 4 are detected as the wear area 9a, as shown in FIG. 5, the wear area 9a is accurately detected. The wear area correction unit 40E transmits to the post-processing unit 41, as the corrected wear area extraction image 50B, the number of corresponding wear areas 9a, length information including the start coordinates and end coordinates in the vertical direction Y of each wear area 9a, and information regarding the image processing execution procedure (or processing pattern) M corresponding to the wear area extraction image.

[0046] In this way, the wear area extraction image generation unit 32 accurately extracts the wear area of the trolley wire 9 by executing the image processing function with the line sensor image 50 as an input according to each of the plurality of image processing execution procedures M by the image processing unit 40 and the post-processing unit 41, and generates a plurality of wear area extraction images 50B corresponding to each of the plurality of image processing execution procedures M.

[0047] Each of the image processing functions included in these contrast adjustment processing unit 40A, background / noise removal processing unit 40B, binarization processing unit 40C, interpolation processing unit 40D, and wear area correction unit 40E is executed according to a processing pattern (image processing execution procedure) M which is the execution order when executing each image processing function registered in the image processing execution procedure storage unit 33.

[0048] FIG. 6 is an explanatory diagram of the image processing execution procedure storage unit of the trolley wire detection device of the present embodiment. In FIG. 6, the image processing execution procedure storage unit 33 includes a process storage unit 33a, a parameter set storage unit 33b, and a process pattern registration unit 33c.

[0049] The processing storage unit 33a stores each of the image processing functions executed in each of the contrast adjustment processing unit 40A, background / noise removal processing unit 40B, binarization processing unit 40C, interpolation processing unit 40D, and wear area correction unit 40E, which have been described above. In the processing storage unit 33a of FIG. 6, X image processing functions PR1 to PRX from PR1 to PRX, which are included in any of the above processing units 40A, 40B, 40C, 40D, and 40E, are stored.

[0050] Further, the parameter set storage unit 33b stores parameter sets that can be used during actual execution of each of the image processing functions PR1 to PRX stored in the processing storage unit 33a. In the parameter set storage unit 33b of FIG. 6, Y types of parameter sets PS1 to PSY, which are used in any one or a plurality of the X image processing functions PR1 to PRX stored in the processing storage unit 33a, are stored.

[0051] The processing pattern registration unit 33c stores a plurality of processing patterns M for executing the above image processing functions when processing the line sensor image 50. In the processing pattern registration unit 33c of FIG. 6, L processing patterns M1 to ML from M1 to ML are stored.

[0052] In FIG. 6, attention is paid to the first processing pattern M1 among these L processing patterns M1 to ML. The processing pattern M1 is registered to sequentially execute Z image processing functions P1 to PZ from the image processing function P1 to the image processing function PZ. Here, the image processing function P1 in the processing pattern M1 refers to the image processing function PR1 in the processing storage unit 33a and the parameter set PS2 in the parameter set storage unit 33b. This indicates that the image processing function P1 in the processing pattern M1 is actually the image processing function PR1 stored in the processing storage unit 33a, which is executed with the parameter set PS2 applied.

[0053] Also, in the image processing function P2 that is executed second in the first processing pattern M1, only the image processing function PRX stored in the processing storage unit 33a is referred to, and the parameter set in the parameter set storage unit 33b is not referred to. Thus, when the image processing functions PR1 to PRX stored in the processing storage unit 33a are registered in the processing pattern M, the parameter sets in the parameter set storage unit 33b do not necessarily have to be associated and referred to.

[0054] Although detailed description is omitted, the processing patterns M2 to ML are also configured by expressing the image processing functions that are executed with reference to the processing storage unit 33a and the parameter set storage unit 33b in the same manner as the processing pattern M1.

[0055] In each of the processing patterns M1 to ML stored in the processing pattern registration unit 33c, the number of image processing functions to be executed is variable length, not fixed length. That is, the number of image processing functions to be executed in each of the processing patterns M1 to ML is basically different from each other.

[0056] Thus, the processing pattern M is configured by registering a plurality of combinations of an image processing function or an image processing function and a parameter set used in the image processing function in an ordered manner. Thus, it can be said that the processing pattern M is the image processing execution procedure M in which the execution order of each image processing function when processing the line sensor image 50 is registered together with the parameter set indicating how each image processing function is actually executed. Hereinafter, the processing pattern M will be referred to as the image processing execution procedure M.

[0057] As described above, each image processing function is classified into a process aimed at any one of contrast adjustment processing, background / noise removal processing, binarization processing, and interpolation processing. In each image processing execution procedure M, the image processing functions are registered so as to be executed in this order of process classification. For example, since it basically makes no sense to execute contrast adjustment after binarization processing, an image processing execution procedure M in which the image processing function included in the binarization processing unit 40C is executed and then the image processing function included in the contrast adjustment processing unit 40A is executed is basically not registered.

[0058] Also, in order to extract the wear area 9a, at least binarization processing needs to be executed. In other words, in the image processing execution procedure M, the image processing function included in the binarization processing unit 40C is executed at least once, and the image processing functions included in the other processing units 40A, 40B, 40D, and 40E are executed as needed, that is, 0 times or more. Thus, the order of image processing functions is registered such that in each of the plurality of image processing execution procedures M, 0 or more contrast adjustment processes, 0 or more background / noise removal processes, 1 or more binarization processes, and 0 or more interpolation processes are executed in this order.

[0059] Furthermore, the order of the image processing functions included in each of the processing units 40A, 40B, 40C, 40D, and 40E may be registered so that each of them is executed a plurality of times. For example, in the interpolation processing unit 40D, it may be possible to construct an image processing execution procedure M such that after connecting the regions divided in the horizontal direction X by closing processing, the regions divided in the vertical direction Y are connected by wavy wear interpolation processing.

[0060] Alternatively, the same image processing function may be registered in the image processing execution procedure M so as to be executed a plurality of times with the same parameter set or with a changed parameter set, continuously or with the execution of other image processing functions included in the same process classification interposed therebetween.

[0061] A plurality of image processing execution procedures M are registered so that all combinations of processing procedures are realized among the image processing functions in each of the processing units 40A, 40B, 40C, 40D, and 40E. In each of these plurality of image processing execution procedures M, the accuracy of extracting the wear area 9a and the suitability for the situation such as the background of the line sensor image 50 also vary widely. For this reason, a priority is associated with and stored in each of the plurality of image processing execution procedures M. For example, it is conceivable to set a high priority for an image processing execution procedure M that is not greatly affected by the luminance value of the background and can always exhibit a wear area extraction accuracy of a certain level or higher. In this way, the priority is set so that the higher the extraction accuracy of the wear area 9a is considered to be, the higher it becomes.

[0062] As described above, the image processing unit 40 executes each image processing execution procedure M, generates a processed image corresponding to each image processing execution procedure M, and transmits it to the post-processing unit 41.

[0063] The post-processing unit 41 receives the processed image from the image processing unit 40, performs edge extraction to extract the wear area 9a, and if a plurality of wear areas 9a are extracted in the processed image, calculates the length in the vertical direction Y of each wear area 9a for each wear area 9a. As already described, in this embodiment, there is a possibility that up to four trolley lines 9 are provided overlapping each other. That is, up to four trolley lines 9 can be imaged in one line sensor image 50. For this reason, when five or more wear areas 9a are extracted, the post-processing unit 41 treats the four areas with the longest lengths as wear areas 9a in order from the longest, and regards the other areas with shorter lengths as not being wear areas 9a. The post-processing unit 41 uses the image after edge extraction as the wear area extraction image 50A, and includes the number of corresponding wear areas 9a, the length information including the start point coordinates and the end point coordinates in the vertical direction Y of each wear area 9a, and information regarding the image processing execution procedure M corresponding to the wear area extraction image, and transmits it to the wear area extraction image evaluation unit 34.

[0064] The wear area extraction image evaluation unit 34 receives the corrected wear area extraction image 50B corresponding to each of the plurality of image processing execution procedures M and related information, calculates the evaluation value of each of the received plurality of wear area extraction images 50B, and selects the wear area extraction image with the highest evaluation value. For each wear area extraction image, in the wear area extraction image generation unit 32, a determination of how many wear areas 9a are imaged in the wear area extraction image is associated and transmitted to the wear area extraction image evaluation unit 34. The wear area extraction image evaluation unit 34 compares the number of wear areas 9a in each wear area extraction image with the number of trolley lines 9 at the actual site where the wear area extraction image is imaged, and calculates the evaluation value of the wear area extraction image.

[0065] As the evaluation value of the wear area extraction image, taking advantage of the fact that the number of trolley lines shown in the image is known from the pre-registered equipment information, for each time series (line) of the image, cost calculation is performed to adopt the wear area extraction image of the processing pattern with the lowest cost (the lower the cost, the higher the evaluation), or for each of the plurality of wear area extraction images, a likelihood evaluation (the higher the likelihood, the higher the evaluation), which is an index indicating that the wear area 9a has features on the image close to the trolley line 9, is used. For the details of the cost evaluation and the likelihood evaluation, refer to Patent No. 7196806.

[0066] Based on the cost evaluation or the likelihood evaluation, the wear area extraction image evaluation unit 34 calculates the centroid coordinates and the wear surface width of each wear area 9a for the wear area extraction image with the highest evaluation value, and transmits them together with the wear area extraction image with the highest evaluation value to the trolley line deviation calculation unit 35 and the trolley line remaining diameter calculation unit 36.

[0067] The trolley line deviation calculation unit 35 receives the wear area extraction image selected as the one with the highest evaluation value, and based on this, calculates the deviation of the trolley line 9. FIG. 7(a) is an explanatory diagram of the principle of calculating the deviation. The deviation D (unit: mm) is expressed by the following formula using the center-of-gravity coordinates d (unit: pixel) of the wear area 9a received from the wear area extraction image evaluation unit 34, the distance hs (mm) from the track 2 to the sensor surface 22c of the line sensor 22, which is a parameter set in advance, the focal length f (mm) of the lens of the line sensor 22, the length s (mm) of the sensor surface 22c of the line sensor 22, the number of pixels p (pixels) of the line sensor 22, and the distance h (mm) from the track 2 to the trolley line 9 calculated by another measuring device other than the trolley line measuring device 20 of the present invention.

[0068]

Equation

[0069] The trolley line deviation calculation unit 35 calculates the deviation of the trolley line 9 as described above, and determines whether the trolley line 9 is located within the allowable deviation range at the actual location corresponding to the wear area extraction image. If the trolley line 9 is not located within the allowable deviation range, it is determined to be abnormal and recorded.

[0070] The trolley line remaining diameter calculation unit 36 receives the wear area extraction image selected as the one with the highest evaluation value, and calculates the remaining diameter of the trolley line 9 based on this. FIG. 7(b) is an explanatory diagram of the calculation principle of the remaining diameter. In the present embodiment, the remaining diameter of the trolley line 9, that is, the height of the trolley line 9 after wear, is calculated as the remaining diameter. The trolley line remaining diameter calculation unit 36 calculates the actual wear surface width W (mm) of the trolley line 9 based on the center-of-gravity coordinates d (pixels) of the wear area 9a received from the wear area extraction image evaluation unit 34 and the wear surface width w (pixels) on the same principle as the trolley line deviation calculation unit 35. The remaining diameter Dr (mm) of the trolley line 9 is expressed by the following formula using this wear surface width W (mm) and the radius of the trolley line 9.

[0071]

Equation

[0072] The trolley wire residual diameter calculation unit 36 calculates the residual diameter of the trolley wire 9 as described above, and determines whether the residual diameter is below the management value. If the residual diameter is below the management value, it is determined to be abnormal, and this is recorded.

[0073] Next, FIG. 8 is a flowchart for explaining the operation of the trolley wire detection method by the trolley wire detection device 20 according to the present embodiment. When the process starts, the line sensor image generation unit 30 arranges, in time series in the vertical direction, for example, 1000 lines of the input image from the line sensor 22, generates a line sensor image 50 as shown in FIG. 3, and transmits it to the wear region extraction image generation unit 32. At the same time, the environment determination unit 31 determines the environment in which the input image was captured based on the line sensor image 50 generated by the line sensor image generation unit 30 (step S10).

[0074] Next, in the wear region extraction image generation unit 32, the contrast adjustment processing unit 40A adjusts the gradation of the line sensor image 50 (step S12), the background / noise removal processing unit 40B removes objects different from the overhead line, such as buildings imaged in the background, and noise (step S14), and the binarization processing unit 40C converts the image with a plurality of gradations into, for example, two gradations of white and black, and extracts only the wear region 9a (step S16).

[0075] Next, the interpolation processing unit 40D interpolates the shape of the wear region 9a by joining together regions that are assumed to be the wear region 9a but are divided into a plurality of regions when the binarization process in the binarization processing unit 40C is executed and an image in which the region assumed to be the wear region 9a is extracted is targeted (step S18).

[0076] Next, the wear area correction unit 40E searches for the boundary between the sliding surface and the non-sliding surface from the position extracted by the wear area extraction process for each line in the vertical direction of the image by raster scanning outward, and corrects the wear area extraction image 50A by setting the position where the luminance value becomes equal to or lower than a preset luminance value as the sliding surface boundary to detect the accurate wear area 9a (step S20).

[0077] Next, the post-processing unit 41 generates, as the corrected wear area extraction image 50B, the number of corresponding wear areas 9a and the length information including the start coordinates and the end coordinates in the vertical direction Y of each wear area 9a, and transmits the information to the wear area extraction image evaluation unit 34 together with the information regarding the image processing execution procedure M corresponding to the wear area extraction image (step S22).

[0078] Next, the wear area extraction image evaluation unit 34 calculates the evaluation value (likelihood evaluation or cost evaluation) of each of the received plurality of wear area extraction images based on the number of corresponding wear areas 9a, the length information including the start coordinates and the end coordinates in the vertical direction Y of each wear area 9a, and the information regarding the image processing execution procedure M corresponding to the wear area extraction image, with the image after edge extraction from the wear area extraction image generation unit 32 as the wear area extraction image (step S24). The above generation and evaluation of the wear area extraction image are executed for all the image processing execution procedures M stored in the image processing execution procedure storage unit 33.

[0079] Next, it is determined whether the generation and evaluation of the wear area extraction image have been executed for all the image processing execution procedures M (step S26). If there is an unexecuted image processing execution procedure M (NO in step S26), the process proceeds to step S10 to continue the process for the unexecuted image processing execution procedure M.

[0080] On the other hand, if there is no unexecuted image processing execution procedure M (YES in step S26), the wear area extraction image evaluation unit 34 determines that the wear area extraction image with the highest evaluation value among all the wear area extraction images is the image in which the wear area 9a is most appropriately extracted (step S28). The wear area extraction image evaluation unit 34 calculates the centroid coordinates and the wear surface width of each wear area 9a for the wear area extraction image with the highest evaluation value, and transmits them together with the wear area extraction image with the highest evaluation value to the trolley line deviation calculation unit 35 and the trolley line remaining diameter calculation unit 36.

[0081] Next, the trolley line deviation calculation unit 35 receives the wear area extraction image selected as the one with the highest evaluation value, and calculates the deviation of the trolley line 9 based on this (step S30). Also, the trolley line remaining diameter calculation unit 36 receives the wear area extraction image selected as the one with the highest evaluation value, and calculates the remaining diameter of the trolley line 9 based on this (step S32). Then, the process ends.

[0082] According to the above-described embodiment, there is provided a trolley wire detection device 20 including a line sensor 22 that images the trolley wire 9, and detecting the trolley wire 9 based on an image captured by the line sensor 22 as an input image. The trolley wire detection device 20 includes: a line sensor image generation unit 30 that generates a line sensor image 50 by arranging input images in time series; an image processing execution procedure storage unit 33 in which a plurality of execution orders are registered as an image processing execution procedure M when sequentially executing a plurality of image processing functions on the line sensor image 50; a wear region extraction image generation unit 32 that extracts a wear region 9a of the trolley wire 9 by executing an image processing function with the line sensor image 50 as an input according to each of the plurality of image processing execution procedures M, corrects the wear region by scanning outward from the extracted wear region of the trolley wire and setting a position where the luminance value becomes equal to or less than a preset luminance value as the boundary of the wear region, and generates a plurality of wear region extraction images corresponding to each of the plurality of image processing execution procedures M; a wear region extraction image evaluation unit 34 that calculates evaluation values of each of the plurality of wear region extraction images and selects the wear region extraction image with the highest evaluation value; a trolley wire deviation calculation unit 35 that calculates the deviation of the trolley wire 9 based on the selected wear region extraction image; and a trolley wire remaining diameter calculation unit 36 that calculates the remaining diameter of the trolley wire 9.

[0083] Also, the trolley wire detection method in the present embodiment is a trolley wire detection method for detecting the trolley wire 9 based on an image captured by a line sensor 22 that captures the trolley wire 9 as an input image. The method includes arranging the input images in a time series to generate a line sensor image 50, and when sequentially executing a plurality of image processing functions on the line sensor image 50, the execution order is registered as a plurality of image processing execution procedures M. According to each of the plurality of image processing execution procedures M, by executing the image processing function with the line sensor image 50 as the input, the worn area 9a of the trolley wire 9 is extracted, and by scanning outward from the extracted worn area 9a and setting the position where the luminance value becomes equal to or less than a preset luminance value as the boundary of the worn area 9a, the worn area 9a is corrected. A plurality of worn area extraction images corresponding to each of the plurality of image processing execution procedures M are generated, the evaluation value of each of the plurality of worn area extraction images is calculated, and the worn area extraction image with the highest evaluation value is selected. Based on the selected worn area extraction image, both the remaining diameter and the deviation of the trolley wire 9 are calculated.

[0084] According to the above configuration, even when a luminance difference occurs in the worn area 9a due to lighting, the state of the worn area 9a of the trolley wire 9, the shooting environment, etc., the worn area 9a of the trolley wire 9 can be accurately detected, the extraction accuracy of the worn area 9a of the trolley wire 9 is improved, and the calculation accuracy of the remaining diameter and the deviation is also improved.

[0085] Also, according to the present embodiment, a worn area extraction image is generated for each of the plurality of image processing execution procedures M, an evaluation value is calculated for each of these worn area extraction images, the worn area extraction image with the highest evaluation value is selected, and based on this, both the remaining diameter and the deviation of the trolley wire 9 are calculated. That is, the number of worn area extraction images corresponding to the number of image processing execution procedures M is generated, and the worn area extraction image most suitable for the evaluation of the remaining diameter and the deviation is selected as the evaluation target for the remaining diameter and the deviation from among them. As a result, among the plurality of registered image processing execution procedures M, the processing result of the image processing execution procedure M suitable for the processing of the line sensor image 50 to be processed is selected as the worn area extraction image.

[0086] In this way, even if an operator does not select a combination of image processing functions suitable for the line sensor image 50 to be processed through trial and error, a wear region extraction image suitable for calculating the remaining diameter and deviation of the trolley line 9 can be generated, and the remaining diameter and deviation can be measured. Therefore, it is possible to reduce the man-hours required for the operator to select and choose an image processing function suitable for the input image.

[0087] Further, according to the present embodiment, a priority is associated with and stored in each of the plurality of image processing execution procedures M. When the same evaluation value is calculated for a plurality of wear region extraction images, the wear region extraction image corresponding to the image processing execution procedure M with a higher priority is selected as the wear region extraction image with the highest evaluation value. Therefore, even when there are a plurality of wear region extraction images having the same evaluation value, the wear region extraction image corresponding to the image processing execution procedure M with a higher priority is selected, so that a wear region extraction image suitable for evaluating the remaining diameter and deviation is likely to be selected.

[0088] Further, according to the present embodiment, the wear region extraction image evaluation unit 34 calculates the number of wear regions 9a for each of the plurality of wear region extraction images, and calculates the evaluation value so that the evaluation value becomes higher when the difference between the number of wear regions 9a and the actual number of trolley lines 9 is small. Therefore, a wear region extraction image in which the actual number of trolley lines 9 matches the number of wear regions 9a is likely to be selected as the wear region extraction image for calculating the remaining diameter and deviation. For this reason, the extraction accuracy of the wear region 9a of the trolley line 9 is improved, and the calculation accuracy of the remaining diameter and deviation is also improved.

[0089] Further, according to the present embodiment, an appropriate image processing execution procedure is selected according to the environment, and a wear area extraction image corresponding thereto is generated. As a result, the calculation accuracy of the remaining diameter and the deviation is improved, and the processing time can be shortened or the processing load can be reduced. Further, since the image processing execution procedure M that does not correspond to the selected environment is not executed, the processing time and memory required for the execution of the trolley wire detection device 20 can be reduced.

[0090] Also, according to the present embodiment, since the wear area extraction image with the highest evaluation value, that is, the trolley wire likelihood, is selected, a wear area extraction image in which the features on the image of the wear area 9a are close to the appearance of the actual trolley wire 9 is likely to be selected as the wear area extraction image for calculating the remaining diameter and the deviation. Therefore, the extraction accuracy of the wear area 9a of the trolley wire 9 is improved, and the calculation accuracy of the remaining diameter and the deviation is also improved.

[0091] In the above embodiment, for example, the post-processing unit 41 has been described as being included in the wear area extraction image generation unit 32 for the sake of explanation, but it is not limited thereto. The post-processing unit 41 may be included in the wear area extraction image evaluation unit 34, or may be configured as an independent function from each of the wear area extraction image generation unit 32 and the wear area extraction image evaluation unit 34. Needless to say, as long as the gist of the present invention is not deviated from, the configuration within the control device 23 may be changed. In the above embodiment, the trolley wire detection device includes both the trolley wire deviation calculation unit 35 and the trolley wire remaining diameter calculation unit 36, and is configured to calculate both the remaining diameter and the deviation of the trolley wire 9, but it is not limited thereto. The trolley wire detection device may be configured to calculate only one of the remaining diameter and the deviation of the trolley wire, for example.

[0092] In addition to this, the present invention is not limited to the above-described embodiments and each modification described with reference to the drawings. As long as the gist of the present invention is not deviated from, it is possible to select and combine the configurations exemplified in the above embodiments and each modification, or to appropriately change to other configurations.

Explanation of Reference Numerals

[0093] 1 Inspection vehicle 9 Trolley line 9a Wear area 9c, 9d Dark area 20 Trolley line detection device 21 Lighting 22 Line sensor 23 Control device 30 Line sensor image generation unit 31 Environment determination unit 32 Wear area extraction image generation unit 33 Image processing execution procedure memory unit 34 Wear area extraction image evaluation unit 35 Trolley line deviation calculation unit (trolley line detection unit) 36 Trolley line remaining diameter calculation unit (trolley line detection unit) 40 Image processing unit 40A Contrast adjustment processing unit 40B Background / noise removal processing unit 40C Binarization processing unit 40D Interpolation processing unit 40E Wear area correction unit 41 Post-processing unit 50 Line sensor image 50A, 50B Wear area extraction image M, M1, M2, M3 Image processing execution procedure

Claims

1. A trolley wire detection device comprising a line sensor for imaging a trolley wire, and detecting the trolley wire based on an image captured by the line sensor as an input image, a line sensor image generation unit that arranges the input images in time series to generate a line sensor image, an image processing execution procedure storage unit in which a plurality of execution orders are registered as image processing execution procedures when sequentially executing a plurality of image processing functions on the line sensor image, By executing the plurality of image processing functions with the line sensor image as an input according to each of the plurality of image processing execution procedures, extracting a worn area of the trolley wire, scanning outward from the extracted worn area of the trolley wire, and setting a position where the luminance value becomes equal to or lower than a preset luminance value as a boundary of the worn area to correct the worn area, and generating a plurality of worn area extraction images corresponding to each of the plurality of image processing execution procedures, a worn area extraction image generation unit, a worn area extraction image evaluation unit that calculates an evaluation value for each of the plurality of worn area extraction images and selects the worn area extraction image with the highest evaluation value, a trolley wire detection unit that calculates either or both of the remaining diameter and deviation of the trolley wire based on the selected worn area extraction image with the highest evaluation value, A trolley wire detection device characterized by comprising.

2. In each of the plurality of image processing execution procedures, 0 or more contrast adjustment processes, 0 or more background / noise removal processes, 1 or more binarization processes, and 0 or more interpolation processes are registered so as to be executed in this order. The trolley wire detection device according to claim 1, characterized by this.

3. Each of the plurality of image processing execution procedures is stored with a priority associated therewith, When the same evaluation value is calculated for a plurality of the wear area extraction images, the wear area extraction image corresponding to the image processing execution procedure with a higher priority is selected as the wear area extraction image with the highest evaluation value. The trolley wire detection device according to claim 1 or 2, characterized in that.

4. For each of the plurality of wear area extraction images, the wear area extraction image evaluation unit calculates the number of the wear areas, and when the difference between the number of the wear areas and the actual number of trolley wires is small, the evaluation value is calculated so that the evaluation value becomes high. The trolley wire detection device according to claim 3, characterized in that.

5. For each of the plurality of wear area extraction images, the wear area extraction image evaluation unit calculates a trolley wire likelihood, which is an index indicating that the wear area has characteristics on an image close to the trolley wire, and when the trolley wire likelihood is high, the evaluation value is calculated so that the evaluation value becomes high. The trolley wire detection device according to claim 3, characterized in that.

6. In the image processing execution procedure storage unit, the image processing execution procedures are registered corresponding to each of a plurality of different environments when the image is captured by the line sensor. The trolley wire detection device further includes an environment determination unit that determines the environment in which the input image was captured based on the line sensor image. The wear area extraction image generation unit generates the wear area extraction image according to the image processing execution procedure corresponding to the determined environment. The trolley wire detection device according to claim 1, characterized in that.

7. A trolley wire detection method for detecting a trolley wire based on an input image which is an image captured by a line sensor for imaging the trolley wire, Generating a line sensor image by arranging the input images in a time series. When sequentially executing a plurality of image processing functions on the line sensor image, the execution order is registered in plural as an image processing execution procedure, and by executing the plurality of image processing functions with the line sensor image as an input in accordance with each of the plurality of image processing execution procedures, an abrasion region of the trolley wire is extracted. Scanning outward from the extracted abrasion region of the trolley wire, a position where the luminance value becomes equal to or less than a preset luminance value is set as a boundary of the abrasion region to correct the abrasion region, and a plurality of abrasion region extraction images corresponding to each of the plurality of image processing execution procedures are generated. Calculating evaluation values of each of the plurality of abrasion region extraction images and selecting the abrasion region extraction image having the highest evaluation value. Calculating either or both of a remaining diameter and a deviation of the trolley wire based on the selected abrasion region extraction image having the highest evaluation value. A trolley wire detection method characterized by including the above.

Images