Crack detection system and crack detection method

Abstract

To provide a crack detection system and a crack detection method allowing all of operators to accurately recognize a position of a crack.SOLUTION: The crack detection system comprises: photographing means 2 for photographing a surface of a structure; crack detection mean 3 for using a learned model for detecting a crack initiation point on the surface of the structure from an image of the surface of the structure, and detecting the crack initiation point from the image photographed by the photographing means 2; image creation means 3 for creating an image in which a symbol is arranged at a position corresponding to the detected crack initiation point; and radiation means 4 for projecting the created image on the surface of the structure, so as to present the crack initiation point with the projected symbol.SELECTED DRAWING: Figure 1

Description

[[Technical Field]]

[0001] The present invention relates to, for example, a crack detection system and a crack detection method for detecting crack occurrence locations in shotcrete. [[Background Art]]

[0002] <Mirror spraying, guidelines> Conventionally, when excavating natural ground in mountain tunnels, "face spalling", in which rock mass peels off and falls from the working face, sometimes occurs during excavation. It has been clarified that once a face spalling disaster occurs, it tends to be a serious disaster: for example, 6% of victims die, 42% take more than one month off work, and 66% take more than four days off work (see, for example, Non-Patent Document 1).

[0003] As a countermeasure against face spalling, an operation called mirror spraying, in which shotcrete is sprayed onto the working face immediately after muck is carried out, is sometimes performed. Mirror spraying has the following effects. <1> Suppresses loosening of the working face. <2> Makes it easier to visually recognize newly generated cracks and deformation of the working face accompanying deformation of the working face. <3> In the case of expansive ground, prevents contact between the ground and air or moisture inside the tunnel, thereby inhibiting expansion of the ground.

[0004] In order to detect signs of collapse, the "Guidelines for Prevention of Face Spalling Disasters at Working Faces in Mountain Tunnel Construction" of Non-Patent Document 1, revised in 2018, requires that a dedicated face observer constantly monitor the condition of the working face. However, there is a problem that humans have a limited ability to visually capture all signs such as small cracks and minor spalling (Problem [1]).

[0005] <Crack detection system> To address the above-mentioned problem [1], an AI-based crack detection system for mirror-sprayed concrete surfaces has been developed to instantly detect cracks in the surface (see, for example, Non-Patent Documents 2 and 3, and Patent Document 1). This system constructs an AI model that can detect cracks in the mirror-sprayed concrete surface in real time using video footage in which cracks are recorded as training data. This system can solve the above-mentioned problem [1].

[0006] On the other hand, as a conventional technique for projecting information related to the tunnel face onto the tunnel face surface, for example, Patent Documents 2 and 3 are known. [Prior art documents] [Non-patent literature]

[0007] [Non-Patent Document 1] "Guidelines for Preventing Skin Collapse Disasters at the Tunnel Face of Mountain Tunnel Construction," published by the Ministry of Health, Labour and Welfare, 2018. [Non-Patent Document 2] "Development of an AI-based crack detection system for sprayed concrete surfaces at tunnel faces," Shunsuke Ariie, Ryo Henmi, and Dota Awaji, 2020 Japan Society of Civil Engineers National Convention, CS15-26, 2020. [Non-Patent Document 3] "Development of an AI-powered detection system for sprayed concrete," Dota Awaji, Shunsuke Ariie, Ryo Henmi, Taiji Mihara, Hiroaki Ihara, Civil Engineering and Construction Technology Presentation Meeting 2020, II-4, 2020. [Patent Documents]

[0008] [Patent Document 1] Japanese Patent Publication No. 2021-184163 [Patent Document 2] Japanese Patent Publication No. 2018-150720 [Patent Document 3] Japanese Patent Publication No. 2022-1732 [Overview of the project] [Problems that the invention aims to solve]

[0009] Incidentally, the conventional crack detection system described above can solve the above-mentioned problems [1]. However, because the location of the detected cracks is viewed on a monitor such as a PC (personal computer), it is difficult for all workers working near the tunnel face to simultaneously recognize the location of the cracks, and there is a possibility of misidentifying the location of the cracks, which could lead to workers mistakenly entering dangerous areas and being involved in accidents such as rockfalls (Problem [2]).

[0010] The present invention has been made in view of the above, and aims to provide a crack detection system and a crack detection method that enable all workers to accurately recognize the location of cracks. [Means for solving the problem]

[0011] To solve the above-mentioned problems and achieve the objective, the crack detection system according to the present invention is a system for detecting and presenting cracks on the surface of a structure, and is characterized by comprising: an imaging means for photographing the surface of the structure; a crack detection means for detecting the crack locations on the surface of the structure from an image of the surface of the structure using a trained model for detecting the crack locations on the surface of the structure from an image of the surface of the structure; an image creation means for creating an image in which symbols are placed at positions corresponding to the detected crack locations; and an illumination means for irradiating the surface of the structure with the created image and presenting the crack locations with the irradiated symbols.

[0012] Furthermore, another crack detection system according to the present invention is characterized in that, in the above-described invention, the imaging means captures the surface of the structure including the image irradiated by the illumination means, and the image creation means adjusts the appearance of the symbols based on the symbols included in the image captured by the imaging means and creates an image in which the symbols are arranged.

[0013] Furthermore, the crack detection method according to the present invention is a method for detecting and presenting cracks on the surface of a structure, and is characterized by comprising the steps of: photographing the surface of the structure; detecting the crack locations from the photographed image using a trained model for detecting crack locations on the surface of the structure from the image of the surface of the structure; creating an image in which symbols are placed at positions corresponding to the detected crack locations; and illuminating the surface of the structure with the created image and presenting the crack locations using the illuminated symbols.

[0014] Furthermore, another crack detection method according to the present invention is characterized in that, in the invention described above, it includes the steps of: photographing the surface of the structure including the irradiated image; and adjusting the appearance of the symbols based on the symbols included in the photographed image and creating an image in which the symbols are arranged. [Effects of the Invention]

[0015] The crack detection system according to the present invention is a system for detecting and presenting cracks on the surface of a structure, comprising: an imaging means for photographing the surface of the structure; a crack detection means for detecting the crack locations on the surface of the structure from an image of the surface of the structure using a trained model; an image creation means for creating an image in which symbols are placed at positions corresponding to the detected crack locations; and an illumination means for illuminating the surface of the structure with the created image and presenting the crack locations using the illuminated symbols. This system has the effect of enabling all workers near the surface of a structure, such as sprayed concrete, to accurately recognize the location of the cracks.

[0016] Furthermore, according to another crack detection system according to the present invention, the imaging means images the surface of the structure including the image irradiated by the irradiation means, and the image creating means adjusts the appearance of the symbol based on the symbol included in the image captured by the imaging means and creates an image with the symbol arranged therein. Therefore, by adjusting the appearance of the symbol to be easily visible, an effect is achieved in that it becomes easier to more accurately recognize crack occurrence locations.

[0017] Furthermore, according to the crack detection method according to the present invention, which is a method for detecting and presenting cracks on the surface of a structure, the method comprises: a step of imaging the surface of the structure; a step of detecting the crack occurrence locations from the captured image using a trained model for detecting crack occurrence locations on the surface of the structure from an image of the surface of the structure; a step of creating an image in which a symbol is arranged at a position corresponding to the detected crack occurrence location; and a step of projecting the created image onto the surface of the structure and presenting the crack occurrence location by the projected symbol. Therefore, an effect is achieved in that all workers located near the surface of a structure such as shotcrete can accurately recognize the position of the crack.

[0018] Furthermore, according to another crack detection method according to the present invention, the method comprises: a step of imaging the surface of the structure including the irradiated image; and a step of adjusting the appearance of the symbol based on the symbol included in the captured image and creating an image with the symbol arranged therein. Therefore, by adjusting the appearance of the symbol to be easily visible, an effect is achieved in that it becomes easier to more accurately recognize crack occurrence locations. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] [Figure 1] Figure 1 is a schematic diagram showing an embodiment of the crack detection system according to the present invention, in which (1) is a front view and (2) is a side view. [Figure 2]Figure 2 is a schematic diagram showing examples of the arrangement of the imaging and illumination equipment, where (1) is arranged on a tripod, (2) is arranged on a steel arch support using magnets, and (3) is arranged on heavy machinery. [Figure 3] Figure 3 is a procedure diagram showing an embodiment of the crack detection method according to the present invention. [Figure 4] Figure 4 is an explanatory diagram of this embodiment, where (1) is a camera image, (2) is the detection result of the crack occurrence location, and (3) is a pixel in the video that was identified as a crack. [Figure 5] Figure 5 is an explanatory diagram of this embodiment, where (1) is image data after plotting the symbols, (2) is the state during image irradiation (before adjusting the size of the symbols), (3) is the result of photographing the sprayed surface being irradiated with the image and extracting the total number of pixels corresponding to the symbols, and (4) is the state during image irradiation (after adjusting the size of the symbols). [Modes for carrying out the invention]

[0020] In response to the above-mentioned problems [1] and [2], the present invention utilizes projection mapping technology in an AI-based crack detection system for mirror-sprayed concrete surfaces, thereby improving the ease of recognizing the locations where cracks have occurred by projecting images onto the crack locations.

[0021] The following describes in detail, with reference to the drawings, an embodiment of the crack detection system and crack detection method according to the present invention, using as an example the case in which an image indicating the location of cracks is projected onto the surface of mirror-sprayed concrete in a mountain tunnel. However, this embodiment does not limit the present invention.

[0022] <Structure> As shown in Figure 1, the crack detection system 10 according to an embodiment of the present invention includes a shooting / irradiation device 1 that photographs the sprayed concrete, analyzes the captured images, creates images to be irradiated onto cracked areas, collapsed areas, and seepage areas, and irradiates the images.

[0023] The shooting and illumination device 1 consists of a camera 2 (shooting means) that takes pictures of the mirror-sprayed concrete (surface of a structure), a crack detection means that analyzes the image taken by the camera 2 and detects the locations of cracks in the mirror-sprayed concrete, and a PC 3 that functions as an image creation means that creates an image to be projected onto the locations of cracks, a projector 4 (illumination means) that projects the image created by the PC 3 onto the object being photographed by the camera 2 (in this case, the mirror-sprayed concrete), and a protective container 5 to protect the camera 2, PC 3, and projector 4 from dust and water.

[0024] The protective container 5 is a rectangular parallelepiped container with sides parallel to the x, y, and z axes of the Cartesian coordinate system. The protective container 5 is provided with a window 6 for taking images with the camera 2 and illuminating with the projector 4. The window 6 is preferably made of transparent acrylic or glass, but is not limited to these materials. Camera 2, PC 3, and projector 4 are connected to a power source. The AI ​​model for detecting crack locations is stored in the recording device on PC 3. To simplify the data processing during image creation, which will be explained later, we assume that the x-axis coordinate values ​​of the focal point 2A of the camera 2 lens and the focal point 4A of the projector 4 lens are the same when the projection direction of the projector 4 is set to the x-axis direction.

[0025] <Placement method> In this embodiment, three methods 1 to 3 are illustrated as ways of arranging the imaging and illumination device 1, but the present invention is not limited thereto.

[0026] First, Method 1 is a method using a tripod. As shown in the one-point perspective view of Figure 2(1), a tripod 12 is set up at a predetermined position on the roadbed 11 inside the tunnel, and the imaging and illumination device 1 is set up on top of it. In addition, heavy machinery 13 is set up at another position on the roadbed 11, and at the tunnel face, sprayed concrete 14 is applied to the surface perpendicular to the direction of excavation, and sprayed concrete 15 and steel arch supports 16 are applied to the other surfaces of the natural ground.

[0027] Method 2 involves using magnets. As shown in the one-point perspective view of Figure 2(2), the imaging and illumination device 1 is installed on the steel arch support structure 16 using magnets 17.

[0028] Method 3 involves installation on heavy machinery. As shown in the side view of Figure 2(3), the imaging and illumination device 1 is installed on the heavy machinery 13. Note that in this figure, the existing sprayed concrete and steel arch supports are omitted from the illustration. Also, although this figure shows the imaging and illumination device 1 installed on the heavy machinery 13, which is a drill jumbo, it is not limited to this and may be installed on other heavy machinery as well.

[0029] <Implementation Procedure> Next, the procedure for implementing the crack detection method using the above-described imaging and irradiation device 1 will be explained with reference to Figure 3.

[0030] As shown in Figure 3, in step S1, first, information on the relative position (Δy, Δz) in the y and z axes of the focal point 4A of the projector lens, relative to the focal point 2A of the camera lens attached to the imaging / illumination device 1, is measured. For measurement, surveying equipment such as a ruler or total station is used. This relative position information is used when creating the image to be projected onto the mirror-sprayed concrete in step S7.

[0031] In the next step, S2, training data is accumulated and an AI model (trained model) is created. For the training data, for example, video footage of cracks in sprayed concrete is used to create the AI ​​model. For the method of creating the AI ​​model, for example, the training method described in Patent Document 1 can be used.

[0032] In the next step S3, the imaging and irradiation device 1 is installed in the arrangement described above at a position where the mirror-sprayed concrete 14 can be photographed and the mirror-sprayed concrete 14 can be irradiated.

[0033] In the next step, S4, the mirror-coated concrete 14 is photographed with camera 2. The following explanation will use the example of the case where the image (video) shown in Figure 4(1) is captured.

[0034] In the next step, S5, the image captured by camera 2 is analyzed by the AI ​​model stored in PC3 to detect the location of cracks, as shown in Figure 4(2). When crack locations are detected, the coordinates (yi,zi) of the pixels considered to be cracks in the image are extracted, as shown in Figure 4(3).

[0035] In the next step, S6, PC3 will create an image to project onto the cracked areas of the mirror-coated concrete 14 at the location captured by camera 2. The specific image creation method is described below.

[0036] First, the coordinates (yi,zi) of the crack location in the image extracted in step S5 are used to calculate the coordinates (Yi,Zi) of the symbol to be plotted as crack information in the image for illumination, using the relative position information (Δy,Δz) in the y and z axes of the focal point 4A of the projector 4 lens, which is based on the focal point 2A of the camera 2 lens measured in step S1, and the following equations (1) and (2).

[0037] Yi=yi+Δy...Equation (1) Zi = zi + Δz ... Equation (2)

[0038] As shown in Figure 5(1), symbols of arbitrary shapes, sizes, and colors, such as squares, triangles, and circles, are plotted on the image data centered at coordinates (Yi, Zi). This process is performed to improve the visibility of crack locations.

[0039] To highlight the color illuminating the cracked area and to reduce the glare felt by the workers, it is preferable, but not limited to, that the image illuminating areas other than the cracked area be a dark color such as gray or black.

[0040] In the next step, S7, the image created by PC3 in step S6 is projected onto the mirror-sprayed concrete 14 using projector 4, as shown in Figure 5(2). This highlights the locations of cracks.

[0041] After projecting an image showing the crack location in step S7, the size of the symbol indicating the crack location is adjusted, and the process proceeds to steps S8 and S9 to make it easier to accurately recognize the crack location.

[0042] In the next step, S8, the mirror-coated concrete 14, onto which the image is projected, is photographed by camera 2.

[0043] In the next step, S9, the size of the symbols plotted on the image for irradiation by PC3 is adjusted. Specifically, as shown in Figures 5(3) and (4), the total number of pixels p1 corresponding to the symbols indicating the crack occurrence location is detected from the image taken in step S8, and the ratio p1 / p2 between this number and the total number of pixels p2 indicating the crack occurrence location detected in step S5 is calculated. If this ratio is smaller than an arbitrarily defined threshold a, the symbols plotted on the image for irradiation are made larger; if it is larger than threshold a, the symbols are made smaller.

[0044] When extracting pixels corresponding to symbols, for example, in step S8, pixels that satisfy all of the following inequalities are extracted from the image captured, the RGB values ​​of the pixels (R1, G1, B1), the RGB values ​​of the symbols plotted on the illumination image (R2, G2, B2), and arbitrarily defined thresholds bR, bG, and bB.

[0045] |R1-R2|≦bR and |G1-G2|≦bG and |B1-B2|≦bB

[0046] In the next step, S10, the image with the adjusted symbol size is projected onto the mirror-coated concrete 14 using projector 4.

[0047] In the next step, S11, a decision is made as to whether or not to perform shooting, video analysis, and irradiation again. This decision may be made by a person or by PC3 based on pre-set criteria. If the result is to perform shooting, video analysis, and irradiation again, the system returns to step S4 (Yes in step S11). Since photographing the sprayed surface while irradiating images showing crack locations may lead to misidentification of the crack locations, it is desirable to temporarily suspend irradiation when performing shooting again.

[0048] If, after irradiation in step S10, further shooting, video analysis, and irradiation are not performed (No in step S11), proceed to step S12 to remove or relocate the equipment.

[0049] By using the above method, it is possible to improve the visibility of cracks in the sprayed concrete of mountain tunnels.

[0050] According to this embodiment, by utilizing an AI model, the location of cracks in the sprayed concrete can be detected instantly. Furthermore, by projecting an image onto the detected crack location with a projector 4, all workers present can accurately recognize the location of the crack, reducing the possibility of workers mistakenly entering dangerous areas and being involved in accidents such as concrete falling. Therefore, it can contribute to improving workplace safety.

[0051] In the above embodiment, cracks occurring in the sprayed concrete of a mountain tunnel were used as an example, but the present invention is not limited to this and can be applied to other uses. For example, it can be similarly applied to the detection of cracks occurring in the sprayed concrete of a slope.

[0052] As described above, the crack detection system according to the present invention is a system for detecting and presenting cracks on the surface of a structure, comprising: an imaging means for photographing the surface of the structure; a crack detection means for detecting the crack locations on the surface of the structure from an image of the surface of the structure using a trained model for detecting the crack locations on the surface of the structure from an image of the surface of the structure; an image creation means for creating an image in which symbols are placed at positions corresponding to the detected crack locations; and an illumination means for illuminating the surface of the structure with the created image and presenting the crack locations using the illuminated symbols. As a result, all workers near the surface of a structure such as sprayed concrete can accurately recognize the location of the cracks.

[0053] Furthermore, according to another crack detection system of the present invention, the imaging means captures the surface of the structure including the image irradiated by the illumination means, and the image creation means adjusts the appearance of the symbols based on the symbols included in the image captured by the imaging means and creates an image in which the symbols are arranged. By adjusting the appearance of the symbols to make them easier to see, it becomes easier to recognize the location of the cracks more accurately.

[0054] Furthermore, the crack detection method according to the present invention is a method for detecting and presenting cracks on the surface of a structure, and includes the steps of: photographing the surface of the structure; detecting the crack locations from the photographed image using a trained model for detecting crack locations on the surface of the structure from the image of the surface of the structure; creating an image in which symbols are placed at the positions corresponding to the detected crack locations; and illuminating the surface of the structure with the created image and presenting the crack locations using the illuminated symbols. As a result, all workers near the surface of a structure such as sprayed concrete can accurately recognize the location of the cracks.

[0055] Furthermore, according to another crack detection method of the present invention, the steps include taking a photograph of the surface of the structure including the irradiated image, and adjusting the appearance of the symbols based on the symbols included in the photographed image to create an image in which the symbols are arranged. By adjusting the appearance of the symbols to make them easier to see, it becomes easier to recognize the location of the cracks more accurately. [Industrial applicability]

[0056] As described above, the crack detection system and crack detection method according to the present invention are useful for detecting cracks occurring on sprayed surfaces such as sprayed concrete, and are particularly suitable for all workers near the sprayed surface to accurately recognize the location of the cracks. [Explanation of Symbols]

[0057] 1. Imaging / Irradiation Equipment 2. Camera (Method of taking pictures) 2A, 4A lens focus 3. PC (crack detection means, image creation means) 4. Projector (irradiation means) 5. Protective container 6 windows 10 Crack detection system 11. Roadbed 12 Tripods 13 Heavy machinery 14. Mirror-coated concrete 15. Sprayed concrete 16 Steel arch support 17 Magnets

Claims

1. A system that detects and displays cracks on the surface of a structure, The system comprises: an imaging means for photographing the surface of the structure; a crack detection means for detecting crack locations on the surface of the structure using a trained model for detecting crack locations on the surface of the structure from an image of the surface of the structure; an image creation means for creating an image in which symbols are placed at positions corresponding to the detected crack locations; and an illumination means for illuminating the surface of the structure with the created image and indicating the crack locations using the illuminated symbols. A crack detection system characterized in that the imaging means captures the surface of the structure including an image irradiated by the illumination means, and the image creation means adjusts the appearance of the symbols based on the symbols included in the image captured by the imaging means and creates an image in which the symbols are arranged.

2. A method for detecting and presenting cracks on the surface of a structure, A crack detection method characterized by comprising the steps of: photographing the surface of the structure; detecting the crack locations from the photographed image using a trained model for detecting crack locations on the surface of the structure from the image of the surface of the structure; creating an image in which symbols are placed at positions corresponding to the detected crack locations; illuminating the surface of the structure with the created image and indicating the crack locations with the illuminated symbols; photographing the surface of the structure including the illuminated image; and adjusting the appearance of the symbols based on the symbols included in the photographed image and creating an image in which the symbols are placed.

Images