Three-dimensional track space evaluation device, three-dimensional track space evaluation method, and program

Abstract

This invention provides a three-dimensional track space evaluation device, a three-dimensional track space evaluation method, and a program that can evaluate the accuracy of a three-dimensional track space based on three-dimensional point cloud data surrounding the track. [Solution] The three-dimensional track space evaluation device 13 includes an image acquisition unit 22 that acquires a three-dimensional image of the area around the track, a calculation unit 24 that calculates a predetermined radius of curvature of the track on the three-dimensional image acquired by the image acquisition unit 22, and an evaluation unit 25 that evaluates the accuracy of the three-dimensional image based on a comparison between the radius of curvature of the track calculated by the calculation unit 24 and a reference radius of curvature of the track.

Description

Technical Field

[0001] The present invention relates to a three-dimensional line space evaluation device, a three-dimensional line space evaluation method, and a program.

Background Art

[0002] As railway facilities installed along a railway line, a special signal light emitter that displays a situation that obstructs the operation of a train with a light signal is known. The special signal light emitter is provided with an emergency stop button. When the emergency stop button is pressed, the special signal light emitter outputs a light signal. Since a train crew member needs to perform a stop operation of the train when confirming the light signal of the special signal light emitter, it is necessary for the train crew member to be able to confirm the light signal of the special signal light emitter when the train is located at a point more than a certain distance in front of the special signal light emitter.

[0003] Therefore, a method (for example, Patent Document 1) has been proposed for generating visibility information of a target facility without involving a lot of manpower based on three-dimensional point cloud data around a railway line.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, when generating three-dimensional point cloud data around a railway line using methods such as those described in Patent Document 1, it is possible that the three-dimensional railway space based on the generated three-dimensional point cloud data may differ from the actual space due to measurement errors by the three-dimensional point cloud measurement device. In order for workers to perform line-of-sight checks using the three-dimensional railway space based on the three-dimensional point cloud data, the three-dimensional railway space must be the same as the actual space. While the three-dimensional railway space based on three-dimensional point cloud data can be used for purposes other than line-of-sight checks, in all cases, the three-dimensional railway space must be the same as the actual space.

[0006] This invention has been made in view of the above problems, and its objective is to provide a three-dimensional track space evaluation device, a three-dimensional track space evaluation method, and a program that can evaluate the accuracy of a three-dimensional track space based on three-dimensional point cloud data around a track. [Means for solving the problem]

[0007] One aspect of the present invention is a three-dimensional track space evaluation device comprising: an image acquisition unit that acquires a three-dimensional image of the area around the track; a calculation unit that calculates a predetermined radius of curvature of the track on the three-dimensional image acquired by the image acquisition unit; and an evaluation unit that evaluates the accuracy of the three-dimensional image based on a comparison of the radius of curvature of the track calculated by the calculation unit with a reference radius of curvature of the track.

[0008] One aspect of the present invention is a three-dimensional track space evaluation device comprising: an image acquisition unit that acquires a three-dimensional image of the area around the track; a calculation unit that calculates a predetermined track gradient on the three-dimensional image acquired by the image acquisition unit; and an evaluation unit that evaluates the accuracy of the three-dimensional image based on a comparison of the track gradient calculated by the calculation unit with a reference track gradient.

[0009] One aspect of the present invention is a three-dimensional track space evaluation device comprising: an image acquisition unit that acquires a three-dimensional image of the area around the track; an extraction unit that extracts at least two predetermined objects from the three-dimensional image acquired by the image acquisition unit; a calculation unit that calculates the distance on the three-dimensional image between the objects extracted by the extraction unit based on the distance in real space per unit pixel on the three-dimensional image; and an evaluation unit that evaluates the accuracy of the three-dimensional image based on a comparison between the distance on the three-dimensional image between the objects calculated by the calculation unit and the distance in real space between the objects.

[0010] One aspect of the present invention is a three-dimensional track space evaluation method, which includes an image acquisition step of acquiring a three-dimensional image of the area around the track; a calculation step of calculating a predetermined radius of curvature of the track on the three-dimensional image acquired in the image acquisition step; and an evaluation step of evaluating the accuracy of the three-dimensional image based on a comparison of the radius of curvature of the track calculated in the calculation step with a reference radius of curvature of the track.

[0011] One aspect of the present invention is a program executable by a computer capable of exchanging information with an input / output device, characterized in that it causes the computer to perform a process that includes: an image acquisition step of acquiring a three-dimensional image of the area around a railway track; a calculation step of calculating a predetermined radius of curvature of the railway track on the three-dimensional image acquired in the image acquisition step; and an evaluation step of evaluating the accuracy of the three-dimensional image based on a comparison of the radius of curvature of the railway track calculated in the calculation step with a reference radius of curvature of the railway track. [Effects of the Invention]

[0012] According to the present invention, it is possible to evaluate the accuracy of a three-dimensional railway track space based on three-dimensional point cloud data of the area surrounding the railway track. [Brief explanation of the drawing]

[0013] [Figure 1] Figure 1 is a diagram illustrating the three-dimensional image generation system 1. [Figure 2] Figure 2 is a diagram illustrating the extracted image frame set. [Figure 3]FIG. 3 is a diagram for explaining the three-dimensional image to be generated. [Figure 4] FIG. 4 is a functional block diagram of the three-dimensional line space evaluation device 13. [Figure 5] FIG. 5 is a diagram for explaining the specification of the reference line T by the extraction unit 23. [Figure 6] FIG. 6 is a flowchart for explaining the accuracy evaluation process of the three-dimensional line space evaluation device 13. [Figure 7] FIG. 7 is a functional block diagram of the three-dimensional line space evaluation device 111. [Figure 8] FIG. 8 is a diagram for explaining the specification of the reference line T by the extraction unit 122. [Figure 9] FIG. 9 is a diagram showing the position of the camera 11 in the vertical direction in the case of the example of FIG. 8. [Figure 10] FIG. 10 is a flowchart for explaining the accuracy evaluation process of the three-dimensional line space evaluation device 111. [Figure 11] FIG. 11 is a functional block diagram of the three-dimensional line space evaluation device 131. [Figure 12] FIG. 12 is a diagram for explaining the specification of the line T as the first reference object and the facility S as the second reference object by the extraction unit 142. [Figure 13] FIG. 13 is a flowchart for explaining the accuracy evaluation process of the three-dimensional line space evaluation device 131. [Figure 14] FIG. 14 is a diagram for explaining the estimation of the coordinate difference between the current frame and the next frame of the camera 11.

Embodiment for Carrying Out the Invention

[0014] Hereinafter, the three-dimensional line space evaluation device 13, the three-dimensional line space evaluation method, and the program according to an embodiment of the present invention will be described with reference to the drawings.

[0015] [First Embodiment: Horizontal Direction] (Three-Dimensional Image Generation System 1) Referring to FIG. 1, the three-dimensional image generation system 1 will be described. The three-dimensional image generation system 1 includes a camera 11, a three-dimensional image generation device 12, a three-dimensional line space evaluation device 13, an input device 14, and an output device 15. These do not necessarily have to be different devices, and may be constituted by a device having a plurality of functions.

[0016] The camera 11 is configured to have the functions of a general camera 11 such as a consumer camcorder. The camera 11 is set at a predetermined position at the front part of a vehicle traveling on a railway line for line-of-sight confirmation, images the front of the vehicle, and supplies the obtained video to the three-dimensional image generation device 12. In the case of this example, it is assumed that position information indicating the horizontal and vertical positions of the camera 11 acquired by, for example, GPS (Global Positioning System) is added to each frame of the video obtained by the camera 11.

[0017] When the three-dimensional image generation device 12 is supplied with the video obtained by imaging the front of the traveling vehicle by the camera 11, it generates a three-dimensional image of the surrounding area of the line based on the video.

[0018] Specifically, the three-dimensional image generation device 12 calculates the moving distance of the camera 11 based on the position information of the camera 11 from each frame of the video obtained by imaging by the camera 11, and extracts, from each frame of the video obtained by imaging by the camera 11, a group of frames in which the moving distance of the camera 11 is about 1 m as shown in FIG. 2.

[0019] Then, the three-dimensional image generation device 12 generates a three-dimensional image, for example, as shown in FIG. 3, based on the extracted group of image frames. The reason for generating the three-dimensional image based on the group of frames in which the moving distance of the camera 11 is about 1 m will be described later.

[0020] The method for generating three-dimensional images by the three-dimensional image generation device 12 may be any method, but specifically, for example, it is possible to use the techniques described in "Junki Yoshino, Hiroyuki Takahashi, Nozomi Nagamine. "Basic study on the construction of a three-dimensional track space using train forward-facing images." Materials of the Institute of Electrical Engineers of Japan. TER= The papers of Technical Meeting on "Transportation and Electric Railway", IEE Japan / Transportation and Electric Railway Research Group [ed.]. Institute of Electrical Engineers of Japan, 2023-05. p1-6." and "Takashi Hongo, Wataru Goda, Ryuta Nakasone, Nozomi Nagamine. "Improvement of accuracy of a three-dimensional track space construction method using mask images." Materials of the Institute of Electrical Engineers of Japan. TER= The papers of Technical Meeting on "Transportation and Electric Railway", IEE Japan / Transportation and Electric Railway Research Group [ed.]. Institute of Electrical Engineers of Japan, 2023-09. p25-30."

[0021] The three-dimensional image generation device 12 supplies the generated three-dimensional image to the three-dimensional track space evaluation device 13 and the output device 15.

[0022] The three-dimensional track space evaluation device 13 evaluates the accuracy of the three-dimensional image supplied from the three-dimensional image generation device 12 and supplies the evaluation result to the output device 15. Details of the three-dimensional track space evaluation device 13 will be described later.

[0023] The input device 14 comprises input devices such as a keyboard, mouse, touch panel, microphone, and an external storage device reader or image input device. The input device 14 receives user operation input and information necessary for processing by the three-dimensional image generation device 12 and the three-dimensional track space evaluation device 13, and supplies it to the three-dimensional image generation device 12 and the three-dimensional track space evaluation device 13.

[0024] The output device 15 includes output devices such as a display, a printer, and a speaker, and outputs three-dimensional images generated by the three-dimensional image generation device 12 and the results of processing by the three-dimensional track space evaluation device 13. Examples of how the three-dimensional images output to the output device 15 can be used will be described later.

[0025] (Three-dimensional track space evaluation device 13) Figure 4 is a functional block diagram of the three-dimensional track space evaluation device 13. The three-dimensional track space evaluation device 13 is composed of an input information acquisition unit 21, an image acquisition unit 22, an extraction unit 23, a calculation unit 24, an evaluation unit 25, and an output processing unit 26.

[0026] The input information acquisition unit 21 supplies the information input from the input device 14 to the extraction unit 23, the evaluation unit 25, and the like.

[0027] The image acquisition unit 22 acquires three-dimensional images and positional information of the camera 11 in each frame from the three-dimensional image generation device 12 and supplies them to the extraction unit 23.

[0028] The extraction unit 23 identifies an image of a reference railway line (hereinafter referred to as the "reference railway line") that will serve as the basis for accuracy evaluation (hereinafter, it will be simply referred to as the "reference railway line") from the three-dimensional image supplied by the image acquisition unit 22. For example, as shown in Figure 5, the extraction unit 23 identifies an image in which a railway line T (hereinafter also referred to as the "reference railway line T") with a radius of curvature of a predetermined size or larger is shown, based on map data or the like. Information for identifying the reference railway line is input to the input device 14 and supplied to the extraction unit 23 via the input information acquisition unit 21.

[0029] The extraction unit 23 acquires the horizontal position of the camera 11 on the image showing the identified railway line, based on the position information of the camera 11 assigned to each frame, and supplies it to the calculation unit 24. The circle P in Figure 5 (hereinafter referred to as position P) indicates the horizontal position of the camera 11. For simplicity, only one reference numeral is used in Figure 5.

[0030] The calculation unit 24 calculates the radius of curvature of the reference track on the three-dimensional image and supplies it to the evaluation unit 25. For example, the calculation unit 24 calculates the radius of curvature of the trajectory of the camera 11's position on the image showing the reference track, which is supplied from the extraction unit 23, as the radius of curvature of the reference track on the three-dimensional image.

[0031] As shown in Figure 5, the horizontal position of the reference track T corresponds to the trajectory of the horizontal position P of the camera 11. Therefore, the radius of curvature of the trajectory of the camera 11's position P in the image showing the reference track T can be calculated as the radius of curvature of the reference track T in the three-dimensional image. For example, in the case of Figure 5, the radius of curvature of the reference track T in the three-dimensional image is calculated to be 104m.

[0032] The evaluation unit 25 evaluates the accuracy of the three-dimensional image based on a comparison between the radius of curvature on the three-dimensional image of the reference track (hereinafter referred to as the evaluation radius of curvature) and the radius of curvature of the reference track in real space (hereinafter referred to as the reference radius of curvature). For example, if the difference between the evaluation radius of curvature and the reference radius of curvature is greater than or equal to a predetermined size, the evaluation unit 25 determines that the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is insufficient. On the other hand, if the difference between the evaluation radius of curvature and the reference radius of curvature is smaller than a predetermined size, the evaluation unit 25 determines that the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is sufficient.

[0033] In the example in Figure 5, if the reference radius of curvature of the reference line T is 105m, the evaluation radius of curvature is 104m as described above, so the relative error is 1m (=105m-104m) / 105m = 0.952%. When the threshold is 0.970%, the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is determined to be sufficient.

[0034] The reference radius of curvature is, for example, the design radius of curvature of the railway line included in the track ledger or map data, and is entered by the user into the input device 14 and supplied to the evaluation unit 25 via the input information acquisition unit 21.

[0035] The evaluation unit 25 outputs the evaluation results to the output device 15 via the output processing unit 26. For example, if the output device 15 indicates that the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is sufficient, the user uses that three-dimensional image to check visibility, etc. If the output device 15 indicates that the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is insufficient, the user generates a new three-dimensional image of the area around the railway tracks without using that image.

[0036] (Accuracy evaluation process) Next, the accuracy evaluation process of the three-dimensional track space evaluation device 13 will be explained with reference to the flowchart in Figure 6.

[0037] In step S1, the image acquisition unit 22 acquires the three-dimensional image and the position information of the camera 11 attached to each frame from the three-dimensional image generation device 12 and supplies it to the extraction unit 23.

[0038] In step S2, the extraction unit 23 identifies an image of a reference line to be used as a basis for accuracy evaluation from the three-dimensional image supplied by the image acquisition unit 22, and acquires the horizontal position of the camera 11 on the image showing the reference line based on the position information of the camera 11 assigned to each frame of the three-dimensional image, and supplies it to the calculation unit 24.

[0039] In step S3, the calculation unit 24 calculates the radius of curvature of the trajectory of the camera 11's position on the image showing the reference track, which is supplied from the extraction unit 23, as the radius of curvature on the three-dimensional image of the reference track (evaluation radius of curvature).

[0040] In step S4, the evaluation unit 25 evaluates the accuracy of the three-dimensional image based on a comparison between the radius of curvature on the three-dimensional image of the reference track (evaluation radius of curvature) and the radius of curvature of the reference track in real space (reference radius of curvature). In step S5, the evaluation result is output to the output device 15 via the output processing unit 26.

[0041] The subsequent processing is now complete.

[0042] In the above description, step S5 was executed immediately after step S4, but steps S1 to S4 may be repeated, and step S5 may be executed after evaluations have been performed for multiple reference lines. In that case, the evaluation unit 25 may perform an overall evaluation that combines the evaluations of the multiple reference lines.

[0043] [Second embodiment: Vertical direction] (Three-dimensional track space evaluation device 111) In the three-dimensional track space evaluation device 13 described above, the evaluation unit 25 evaluated the accuracy of the three-dimensional image based on a comparison between the radius of curvature on the three-dimensional image of the reference track (evaluation radius of curvature) and the radius of curvature of the reference track in real space (reference radius of curvature). In other words, the three-dimensional track space evaluation device 13 evaluated the accuracy in the horizontal direction of the three-dimensional space, but it may also evaluate the accuracy in the vertical direction of the three-dimensional space.

[0044] Figure 7 is a functional block diagram of a three-dimensional line space evaluation device 111 that evaluates accuracy in the vertical direction of three-dimensional space. The three-dimensional line space evaluation device 111 is composed of an input information acquisition unit 121, an image acquisition unit 22, an extraction unit 122, a calculation unit 123, an evaluation unit 124, and an output processing unit 26. Components having the same functions as the three-dimensional line space evaluation device 13 in Figure 4 are denoted by the same reference numerals as in Figure 4, and their descriptions are omitted.

[0045] The input information acquisition unit 121 supplies the information input from the input device 14 to the extraction unit 122, the evaluation unit 124, and the like.

[0046] The extraction unit 122 identifies images of reference lines from the three-dimensional images supplied by the image acquisition unit 22. For example, as shown in Figure 8, the extraction unit 122 identifies images containing lines T with a gradient greater than or equal to a predetermined size, based on map data. Information for identifying reference lines is input to the input device 14 and supplied to the extraction unit 122 via the input information acquisition unit 121.

[0047] The extraction unit 122 acquires the vertical position of the camera 11 on the image showing the identified railway line, based on the vertical position information of the camera 11 in each frame, and supplies it to the calculation unit 123. The circle P in Figure 8 (hereinafter referred to as position P) indicates the vertical position of the camera 11. For simplicity, only one reference numeral is used in Figure 8.

[0048] The calculation unit 123 calculates the gradient on the three-dimensional image of the reference track and supplies it to the evaluation unit 124. For example, the calculation unit 123 calculates the gradient of the trajectory of the camera 11's position on the image showing the reference track, which is supplied from the extraction unit 122, as the gradient on the three-dimensional image of the reference track.

[0049] As shown in Figure 8, the vertical position of the reference track T corresponds to the trajectory of the vertical position P of the camera 11. Figure 9 shows the vertical position of the camera 11 in the example of Figure 8, which corresponds to the height of the reference track T in the three-dimensional image. Therefore, the gradient of the trajectory of the camera 11's position in the image showing the reference track T can be calculated as the gradient of the reference track in the three-dimensional image. For example, in the case of Figure 8, the gradient of the track T in the three-dimensional image is calculated to be 1.40%.

[0050] The calculation unit 123 may also calculate the gradient based on the difference in height between two points in the railway line extracted by the extraction unit 122.

[0051] The evaluation unit 124 evaluates the accuracy of the three-dimensional image based on a comparison between the gradient on the three-dimensional image of the reference track (hereinafter referred to as the evaluation gradient) and the gradient of the reference track in real space (hereinafter referred to as the reference gradient). For example, if the difference between the evaluation gradient and the reference gradient is greater than a predetermined size, the evaluation unit 124 determines that the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is insufficient. On the other hand, if the difference between the evaluation gradient and the reference gradient is smaller than a predetermined size, the evaluation unit 124 determines that the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is sufficient.

[0052] In the example in Figure 8, if the reference gradient of the reference line T is 1.40%, the evaluation gradient is 1.35% as described above, so the relative error is 0.05% (=1.40%-1.35%) / 1.40% = 3.57%. If the threshold is 2.00%, the accuracy of the evaluation gradient and the three-dimensional image generated by the three-dimensional image generation device 12 is determined to be insufficient.

[0053] The reference gradient is, for example, the design gradient of the railway line included in the track register or map data, and is input by the user into the input device 14 and supplied to the evaluation unit 124 via the input information acquisition unit 21.

[0054] The evaluation unit 124, like the evaluation unit 25 of the three-dimensional track space evaluation device 13, outputs the evaluation results to the output device 15 via the output processing unit 26.

[0055] (Accuracy evaluation process) Next, the accuracy evaluation process by the three-dimensional track space evaluation device 111 will be explained with reference to the flowchart in Figure 10.

[0056] Steps S11 and S12 involve the same processing as in steps S1 and S2 of Figure 6, so the explanation is omitted.

[0057] In step S13, the calculation unit 123 calculates the gradient of the trajectory of the camera 11's position on the image showing the reference track, supplied by the extraction unit 122, as the evaluation gradient.

[0058] In step S14, the evaluation unit 124 evaluates the accuracy of the three-dimensional image based on a comparison between the gradient on the three-dimensional image of the reference track (evaluation gradient) and the gradient of the reference track in real space (reference gradient). In step S15, the evaluation result is output to the output device 15 via the output processing unit 26.

[0059] The subsequent processing is now complete.

[0060] In the above description, step S15 was executed immediately after step S14, but steps S11 to S14 may be repeated, and step S15 may be executed after evaluations have been performed on multiple reference lines. In that case, the evaluation unit 124 may perform an overall evaluation that combines the evaluations of the multiple reference lines.

[0061] In the above, the evaluation of horizontal accuracy in three-dimensional space (three-dimensional track space evaluation device 13) and the evaluation of vertical accuracy in three-dimensional space (three-dimensional track space evaluation device 111) were explained separately. However, it is also possible to evaluate the accuracy of three-dimensional space by combining the evaluation of horizontal accuracy and vertical accuracy.

[0062] [Third embodiment: Distance from the object] (Three-dimensional track space evaluation device 131) The three-dimensional track space evaluation devices 13 and 111 described above evaluated the accuracy of the three-dimensional space using the radius of curvature and gradient of the track, but it is also possible to use equipment and other facilities set up near the track, in addition to the track itself.

[0063] Figure 11 is a functional block diagram of the three-dimensional track space evaluation device 131 that utilizes equipment and other facilities set up near the railway tracks.

[0064] The three-dimensional track space evaluation device 131 is composed of an input information acquisition unit 141, an image acquisition unit 22, an extraction unit 142, a calculation unit 143, an evaluation unit 144, and an output processing unit 26. Components having the same functions as the three-dimensional track space evaluation device 13 in Figure 4 are denoted by the same reference numerals as in Figure 4, and their descriptions are omitted.

[0065] The input information acquisition unit 141 supplies the information input from the input device 14 to the extraction unit 142, the evaluation unit 144, and the like.

[0066] The extraction unit 142 identifies images of two objects that serve as the basis for accuracy evaluation (hereinafter referred to as the first reference object and the second reference object, respectively) from the three-dimensional image supplied from the image acquisition unit 22. For example, as shown in Figure 12, the extraction unit 142 identifies images showing the railway line T as the first reference object and the equipment S as the second reference object, which are located at a predetermined distance, based on map data or the like. Information for identifying the first and second reference objects is input to the input device 14 and supplied to the extraction unit 142 via the input information acquisition unit 21.

[0067] The calculation unit 143 calculates the shortest distance on the three-dimensional image between the first reference object and the second reference object identified by the extraction unit 142 and supplies it to the evaluation unit 144. For example, as shown in Figure 12, the calculation unit 143 determines the number of pixels for the shortest distance between the track T, which is the first reference object, and the equipment S, which is the second reference object, from the three-dimensional image. Then, the calculation unit 143 calculates the distance in real space (2200 mm) of the shortest distance L on the three-dimensional image between the track T and the equipment S, based on the size (number of pixels) of the reference size (e.g., the track gauge width L of the track T) in three-dimensional space and its size in real space. For example, the calculation unit 143 determines the distance per pixel based on the size of the track gauge width L of the track T in three-dimensional space and its size in real space, and calculates the distance L' in real space from the shortest distance L' (number of pixels) on the three-dimensional image between the track T and the equipment S.

[0068] The evaluation unit 144 evaluates the accuracy of the three-dimensional image based on a comparison between the shortest distance on the three-dimensional image between the first reference object and the second reference object (hereinafter referred to as the evaluation distance) and the shortest distance in real space between the first reference object and the second reference object (hereinafter referred to as the reference distance). For example, if the difference between the evaluation distance and the reference distance is greater than a predetermined size, the evaluation unit 144 determines that the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is insufficient. On the other hand, if the difference between the evaluation distance and the reference distance is less than a predetermined size, the evaluation unit 144 determines that the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is sufficient.

[0069] For example, in the example in Figure 12, if the shortest distance in real space between track T and equipment S (reference distance) is 2100 mm, and the shortest distance in the three-dimensional image between track T and equipment S (evaluation distance) is 2200 mm, the relative error is 100 mm (= 2200 mm - 2100 m) / 2100 m = 4.76%. If the threshold is 2.50%, the accuracy of the three-dimensional image generated by the three-dimensional image generation device 12 is determined to be insufficient.

[0070] The reference distance is, for example, the positional relationship in the design included in the track register or map data, and is entered by the user into the input device 14 and supplied to the evaluation unit 144 via the input information acquisition unit 141.

[0071] The evaluation unit 144 outputs the evaluation results to the output device 15 via the output processing unit 26.

[0072] (Accuracy evaluation process) Next, the accuracy evaluation process of the three-dimensional track space evaluation device 131 will be explained with reference to the flowchart in Figure 13.

[0073] In step S21, the image acquisition unit 22 acquires the three-dimensional image and the position information of the camera 11 attached to each frame from the three-dimensional image generation device 12 and supplies it to the extraction unit 142.

[0074] In step S22, the extraction unit 142 identifies a first reference object and a second reference object, which will serve as the basis for accuracy evaluation, from the three-dimensional image supplied from the image acquisition unit 22, and supplies them to the calculation unit 143.

[0075] In step S23, the calculation unit 143 calculates the shortest distance (evaluation distance) on the three-dimensional image between the first reference object and the second reference object and supplies it to the evaluation unit 144. For example, the calculation unit 143 determines the distance per pixel based on the size of the reference object (in the case of Figure 12, the track gauge width L of the railway line T) in three-dimensional space and its size in real space, and calculates the distance on the three-dimensional image from the size (number of pixels) of the first reference object and the second reference object as the evaluation distance.

[0076] In step S24, the evaluation unit 144 evaluates the accuracy of the three-dimensional image based on a comparison between the shortest distance on the three-dimensional image between the first reference object and the second reference object (evaluation distance) and the shortest distance in real space between the first reference object and the second reference object (reference distance). In step S5, the evaluation result is output to the output device 15 via the output processing unit 26.

[0077] The subsequent processing is now complete.

[0078] In the above description, step S25 was executed immediately after step S24, but steps S21 to S24 may be repeated, and step S25 may be executed after evaluations have been performed on multiple reference lines. In that case, the evaluation unit 144 may perform an overall evaluation that combines the evaluations of the multiple reference lines.

[0079] Furthermore, the three-dimensional space accuracy evaluation process described above may be combined with accuracy evaluation using railway lines (three-dimensional railway space evaluation device 13, 111).

[0080] [Other examples] (Reason for generating a 3D image based on a set of frames where camera 11 traveled approximately 1 meter)

[0081] As described above, the three-dimensional image generation device 12 extracted a group of frames from each frame of the video captured by the camera 11, where the movement distance of the camera 11, i.e., the movement distance of the vehicle, was approximately 1 meter. The reason for this is explained below.

[0082] Generally, when constructing a 3D space, it is preferable that the overlap rate between each frame be 65% or higher in order to stably match feature points. Therefore, when constructing a 3D track space in a railway, it is also preferable that the overlap rate be 65% or higher. As a result of the experiment, the overlap rate was generally 65% ​​or higher between frames where the camera 11 traveled approximately 1m, so a group of frames where the camera 11 traveled approximately 1m was extracted.

[0083] (Other methods for detecting the position of camera 11) As described above, each frame of the video obtained by camera 11 is accompanied by location information of camera 11, for example, obtained by GPS, and the three-dimensional image generation device 12 detects the position of camera 11 based on that location information.

[0084] However, the three-dimensional image generation device 12 can also detect the position of the camera 11 by a method described later.

[0085] For example, the three-dimensional image generation device 12 first uses a feature point detection algorithm to detect characteristic points from a predetermined frame and the next frame. Next, the three-dimensional image generation device 12 assigns a descriptor with unique characteristics to each pixel of the detected feature point. The three-dimensional image generation device 12 evaluates the similarity of the descriptors between the feature point detected from one composite image and the feature point detected from the other composite image, and matches the feature points with the highest similarity. Then, the three-dimensional image generation device 12 detects the amount of movement of the matched feature point between the predetermined frame and the next frame, and, as shown in Figure 14, estimates the coordinate difference with the next frame of the camera 11 based on that amount of movement.

[0086] (Other examples of calculating radius of curvature and gradient) As described above, the calculation units 24 and 123 calculated the radius of curvature and gradient of the railway track in the three-dimensional image based on the position information of the camera 11, but they may also be calculated based on other methods. For example, the calculation units 24 and 123 may be calculated based on known image recognition techniques.

[0087] [Examples of using 3D images] For example, a user can check the line of sight to the target equipment by reviewing the three-dimensional image generated by the three-dimensional image generation device 12. For example, based on an image generated from a predetermined viewpoint position towards the target equipment based on the three-dimensional image, the user can determine whether or not the object to be checked for line of sight is visible. Note that the line of sight check may be performed manually or automatically using image recognition technology or the like.

[0088] Furthermore, for example, users can use three-dimensional images to measure the size of equipment, for instance, to measure the distance of wiring required for equipment installed around railway tracks.

[0089] Furthermore, access to areas around railway tracks is generally difficult, and the times when access is permitted are limited. Therefore, by using three-dimensional images for training and other purposes, training can be conducted more easily and without being constrained by time.

[0090] [Supplementary explanation of the embodiment] The embodiments described above are all preferred examples of the present invention. The numerical values, components, arrangement positions and connection configurations of the components, and processing order in the flowcharts shown in the embodiments are examples only and are not intended to limit the present invention. Furthermore, the figures are not necessarily strictly illustrative.

[0091] The series of processes described above can be executed by hardware or by software. When the series of processes are executed by software, the programs that make up the software are installed from a program storage medium onto a computer that is built into dedicated hardware, or onto an information processing device such as a general-purpose personal computer that can perform various functions by installing various programs.

[0092] The programs executed by the computer may be programs that are processed chronologically in the order described herein, or they may be programs that are processed in parallel or at necessary times, such as when a call is made.

[0093] Furthermore, the embodiments of the present invention are not limited to those described above, and various modifications are possible without departing from the spirit of the invention.

[0094] [Note] The contents described in some of the embodiments above can be understood, for example, as follows:

[0095] (1)Horizontal direction The three-dimensional track space evaluation device 13 (Figure 4) described above is An image acquisition unit 22 acquires a three-dimensional image of the area around the railway tracks, A calculation unit 24 calculates a predetermined radius of curvature of the railway line on the three-dimensional image acquired by the image acquisition unit 22, An evaluation unit 25 evaluates the accuracy of the three-dimensional image based on a comparison between the radius of curvature of the track calculated by the calculation unit 24 and the reference radius of curvature of the track. It is equipped with.

[0096] With this configuration, the three-dimensional track space evaluation device 13 can evaluate the accuracy of the three-dimensional image in the horizontal direction. Furthermore, the evaluation unit 25 evaluates the accuracy of the three-dimensional image based on a comparison with a reference value of the radius of curvature of the track, which is easily obtained from track ledgers, etc., thus facilitating accuracy evaluation.

[0097] (2) Vertical direction The three-dimensional track space evaluation device 111 (Figure 7) described above is, An image acquisition unit 22 acquires a three-dimensional image of the area around the railway tracks, The image acquisition unit 22 has a calculation unit 123 that calculates a predetermined gradient of the railway line on the three-dimensional image, Based on a comparison between the gradient of the track calculated by the calculation unit 123 and the reference gradient of the track, an evaluation unit 124 evaluates the accuracy of the three-dimensional image. It is equipped with.

[0098] With this configuration, the three-dimensional track space evaluation device 111 can evaluate the accuracy of the three-dimensional image in the vertical direction. Furthermore, the evaluation unit 124 evaluates the accuracy of the three-dimensional image based on a comparison with a reference value of the track gradient that is easily obtained from map data, etc., thus facilitating accuracy evaluation.

[0099] (3) Use of the object The three-dimensional track space evaluation device 131 (Figure 11) described above is, An image acquisition unit 22 acquires a three-dimensional image of the area around the railway tracks, The extraction unit 142 extracts at least two predetermined objects from the three-dimensional image acquired by the image acquisition unit 22, A calculation unit 143 calculates the distance in the three-dimensional image between the objects extracted by the extraction unit 142 based on the distance in real space per unit pixel in the three-dimensional image, An evaluation unit 144 evaluates the accuracy of the three-dimensional image based on a comparison between the distance between the objects on the three-dimensional image calculated by the calculation unit 143 and the distance between the objects in real space, It is equipped with.

[0100] This configuration allows the three-dimensional track space evaluation device 131 to evaluate the accuracy of the three-dimensional image. Furthermore, the evaluation unit 144 evaluates the accuracy of the three-dimensional image based on a comparison with reference values ​​of distances between objects that can be easily obtained from track ledgers, map data, etc., thus facilitating accuracy evaluation.

[0101] (4) Radius of curvature of the route based on camera position, etc. Furthermore, the three-dimensional track space evaluation devices 13,111 are The three-dimensional image of the area around the railway tracks is created based on images from a camera 11 mounted on a vehicle traveling on the railway tracks. The image acquisition unit 22 acquires the position information of the camera 11 for each frame of the video, The calculation units 24 and 123 calculate at least one of the radius of curvature of the track and the gradient of the track based on the position information of the camera 11.

[0102] The calculation units 24 and 123 can calculate the radius of curvature and gradient based on the position information of the camera 11, without having to recognize the image and calculate the gradient, thus making the calculation of the radius of curvature and gradient easier. [Explanation of Symbols]

[0103] 1. Three-dimensional image generation system 11 Cameras 12 Three-dimensional image generation device 13, 111, 131 Three-dimensional track space evaluation device 14 Input devices 15 Output device 21, 121, 141 Input Information Acquisition Unit 22 Image acquisition unit 23, 122, 142 Extraction part 24, 123, 143 Calculation Unit 25, 124, 144 Evaluation Department 26 Output Processing Unit

Claims

1. An image acquisition unit that acquires a three-dimensional image of the area around the railway tracks, A calculation unit calculates a predetermined radius of curvature of the railway line on the three-dimensional image acquired by the image acquisition unit, An evaluation unit evaluates the accuracy of the three-dimensional image based on a comparison between the radius of curvature of the track calculated by the calculation unit and the reference radius of curvature of the track. A three-dimensional track space evaluation device characterized by comprising the following features.

2. An image acquisition unit that acquires a three-dimensional image of the area around the railway tracks, A calculation unit calculates a predetermined gradient of the railway line on the three-dimensional image acquired by the image acquisition unit, An evaluation unit evaluates the accuracy of the three-dimensional image based on a comparison between the gradient of the track calculated by the calculation unit and the reference gradient of the track. A three-dimensional track space evaluation device characterized by comprising the following features.

3. An image acquisition unit that acquires a three-dimensional image of the area around the railway tracks, An extraction unit extracts at least two predetermined objects from the three-dimensional image acquired by the image acquisition unit, A calculation unit calculates the distance in the three-dimensional image between the objects extracted by the extraction unit based on the distance in real space per unit pixel in the three-dimensional image, An evaluation unit evaluates the accuracy of the three-dimensional image based on a comparison between the distance between the objects in the three-dimensional image calculated by the calculation unit and the distance between the objects in real space. A three-dimensional track space evaluation device characterized by comprising the following features.

4. A three-dimensional line space evaluation device according to claim 1 or 2, The three-dimensional image of the area around the railway tracks is created based on images from a camera mounted on a vehicle traveling on the railway tracks. The image acquisition unit acquires the position information of the camera 11 for each frame of the video. The calculation unit calculates at least one of the radius of curvature of the track and the gradient of the track based on the position information of the camera 11. A three-dimensional track space evaluation device characterized by the following features.

5. Image acquisition step to obtain a three-dimensional image of the area around the railway tracks, A calculation step for calculating a predetermined radius of curvature of the railway line on the three-dimensional image acquired in the image acquisition step, An evaluation step to evaluate the accuracy of the three-dimensional image based on a comparison between the radius of curvature of the track calculated in the calculation step and the reference radius of curvature of the track, A method for evaluating a three-dimensional railway space, characterized by including the following:

6. In a program that can be executed by a computer capable of exchanging information with an input / output device, Image acquisition step to obtain a three-dimensional image of the area around the railway tracks, A calculation step for calculating a predetermined radius of curvature of the railway line on the three-dimensional image acquired in the image acquisition step, An evaluation step to evaluate the accuracy of the three-dimensional image based on a comparison between the radius of curvature of the track calculated in the calculation step and the reference radius of curvature of the track, A program that causes a computer to perform a process, characterized by including [a certain element].

Images