A method for identifying cracks on the tunnel face and an image extraction device

Abstract

This invention discloses a method and image extraction device for identifying cracks at the tunnel face, relating to the field of tunnel face crack detection technology. The method involves selecting a safe and flat area in front of the tunnel face and setting up a central shooting point and surrounding shooting points. The central shooting point is located on the tunnel axis. A binocular camera is used to capture images at different heights from each shooting point. The binocular camera is equipped with an infrared range sensor and an electronic compass. The electronic compass detects the shooting direction to ensure that each shooting direction is along the tunnel axis. The infrared range sensor detects the actual distance between the binocular camera and the tunnel face during each shooting. This actual distance allows for analysis and calculation of each captured image, converting it into the true width and height of each image frame. Using this as a reference, a panoramic image of the tunnel face can be acquired. The panoramic image of the tunnel face has a wide recognition range and clear detailed features, which helps ensure the accuracy of crack identification.

Description

Technical Field

[0001] This invention relates to the field of tunnel face crack detection technology, specifically to a tunnel face crack identification method and image extraction device. Background Technology

[0002] In the construction of highway and railway tunnels, hydraulic tunnels, underground mines, etc., image extraction and crack identification of the working face are important means to assess the construction status and quality. Among them, the measurement and statistics of cracks at the working face constitute the basis for rock mass quality evaluation and surrounding rock classification. However, crack identification is based on the extracted working face images, and its accuracy is directly affected by the detailed features of the extracted working face images and the identification range.

[0003] Currently, in many engineering projects, when acquiring images of the tunnel face, specialized photographers typically capture panoramic images of the face directly, followed by further image processing. However, images obtained in this way lack clear detail, are inaccurate due to the shooting angle, and struggle to accurately determine the true dimensions of the images, thus making it difficult to determine the actual size of cracks at the tunnel face. Current solutions often involve manually calibrating specific landmarks or objects in the image to estimate the face size and calculate crack lengths. This method not only increases workload but can also introduce significant errors during calibration, leading to inaccurate measurement and calculation of crack parameters. Furthermore, the harsh environmental conditions of underground engineering, with a lack of flat areas for photography, limit the flexibility of tunnel face photography, significantly impacting image acquisition during construction.

[0004] Existing technologies also include relatively intelligent image extraction devices. Their basic working principle involves setting a distance in front of the tunnel face and moving the imaging device laterally (horizontally and perpendicular to the tunnel axis) and longitudinally to capture images of different local locations on the tunnel face. These images are then stitched together to form a panoramic image of the tunnel face with clear details. This method overcomes the shortcomings of "inaccurate measurement" and "inaccurate calculation" in fracture identification. However, as mentioned earlier, due to the severe environmental conditions of underground engineering, it is difficult to find large, flat areas in front of the tunnel face for laterally moving the imaging device. Existing technologies typically use additional beams and rails to achieve lateral sliding imaging. This makes the overall structure of the image extraction device relatively complex, large in size, and difficult to transport. Each time an image is taken, a suitable location must be selected to install the beams and rails, making installation cumbersome and inflexible. Furthermore, it is difficult to guarantee the lateral and horizontal positional accuracy of the beams and rails during installation, and this installation error can also contribute to inaccurate image extraction from the tunnel face. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for identifying cracks on the tunnel face and an image extraction device. This method can form a panoramic image with a wide recognition range and clear details during image extraction. The device has a simple structure, is portable, and can achieve flexible shooting.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] A method for identifying cracks at the tunnel face includes the following steps:

[0008] S1. A central shooting point and multiple peripheral shooting points are set in front of the working face. The central shooting point and peripheral shooting points are set in flat and safe areas according to the actual situation. The central shooting point is set on the tunnel axis in front of the working face.

[0009] S2. Take pictures at different heights at the central shooting point using a binocular camera. During shooting, the binocular camera is kept horizontal and facing the working face. Record the shooting direction of the binocular camera during shooting. Import each picture into the computer and record the actual distance between the binocular camera and the working face, the focal length of the binocular camera and the magnification of the binocular camera during each shooting.

[0010] S3. Take pictures at different heights at each of the surrounding shooting points using a binocular camera. During shooting, the binocular camera is kept horizontal and the shooting direction is consistent with the direction when shooting at the central shooting point. Import the captured pictures into the computer and record the actual distance between the binocular camera and the working face, the focal length of the binocular camera and the magnification of the binocular camera at each shooting.

[0011] S4. The computer analyzes and calculates the imported data, and calculates the actual width and height of the image based on the actual distance between the binocular camera and the working face corresponding to each image, the focal length and magnification of the binocular camera, and the physical parameters of the binocular camera.

[0012] S5. The computer uses an image stitching toolbox to scale the image according to the actual width and height of the captured image, and then completes the stitching of the panoramic image of the working face to obtain the panoramic image of the working face and calculates the actual width and height of the panoramic image of the working face.

[0013] Furthermore, after step S5, the following steps are also performed:

[0014] S6. The computer identifies and extracts crack trace information in the panoramic image of the tunnel face through the U-net network model, thereby realizing the vectorization of the crack trace image;

[0015] S7. The computer imports the vectorized image into the fracture orientation extraction system to obtain information on fracture length, fracture dip angle and fracture inclination.

[0016] Specifically, the formulas for calculating the actual width and actual height of the image in step S4 are as follows:

[0017] When N≤N0:

[0018] ,

[0019] ,

[0020] When N > N0:

[0021] ,

[0022] ,

[0023] Wherein, D is the actual width of the image; H is the actual height of the image; f is the focal length of the stereo camera when taking the image; f0 is the critical focal length between optical zoom and digital zoom of the stereo camera, which is a physical parameter of the stereo camera; N is the magnification of the stereo camera when taking the image; N0 is the critical magnification between optical zoom and digital zoom of the stereo camera, which is a physical parameter of the stereo camera; L is the actual distance between the stereo camera and the working face when taking the image; B is the center line of the two lens sensors in the stereo camera, which is a physical parameter of the stereo camera; h is the height of the lens sensors in the stereo camera, which is a physical parameter of the stereo camera; and d is the width of the lens sensors in the stereo camera, which is a physical parameter of the stereo camera.

[0024] Specifically, obtaining the fracture length information in step S7 involves importing the vectorized image into the fracture orientation extraction system, extracting the coordinates of each pixel in the image, using the 8-neighborhood tracking method to find the fracture endpoints and track the fracture path to extract the skeleton of the entire fracture; then calculating the distance between consecutive points on the fracture path according to the Euclidean distance formula, and accumulating them to obtain the pixel length of the entire fracture; finally, converting the pixel length of the fracture into the actual length based on the true width and true height corresponding to the panoramic image of the tunnel face.

[0025] Specifically, obtaining the fracture dip angle information in step S7 refers to importing the vectorized image into the fracture orientation extraction system, extracting the coordinates of each pixel in the image, performing linear fitting on all pixel coordinates of a single fracture skeleton using a linear regression algorithm combined with a loop function to obtain the slope of the straight line, and converting the slope into an angle to obtain the fracture apparent dip angle; obtaining the fracture dip direction in step S7 refers to calculating the fracture dip direction based on the shooting direction of the binocular camera and the fracture apparent dip angle.

[0026] A face-of-working space image extraction device is used to implement the aforementioned face-of-working space crack identification method. It includes a computer and a binocular camera. The binocular camera is equipped with an infrared range sensor and an electronic compass. The infrared range sensor is installed at the front end of the binocular camera and positioned at the midpoint of the center line between the two lens sensors of the binocular camera. The infrared range sensor is used to detect the distance between the object being photographed in front of the binocular camera and the camera itself. The electronic compass is used to detect the shooting direction of the binocular camera. The binocular camera, infrared sensor, and electronic compass are all electrically connected to the computer.

[0027] Furthermore, it also includes a leveling base, which comprises a base plate and a top plate, the base plate and the top plate being connected by three anchor screws. A level is installed on the top plate, and the binocular camera is mounted on the top plate, with the shooting direction of the binocular camera parallel to the top plate.

[0028] Furthermore, it also includes a rotary base, which is connected to the top plate, and the binocular camera is rotatably connected to the rotary base.

[0029] Furthermore, it also includes a screw jack, the housing of which is fixedly connected to the top plate, the screw of which is perpendicular to the top plate, and the rotary seat is fixedly installed on the top end of the screw of which is the screw jack.

[0030] Furthermore, it also includes a tripod, the tripod head of which is detachably connected to the base plate.

[0031] The beneficial effects of this invention are:

[0032] This invention provides a method for identifying fractures at the tunnel face, enabling the acquisition of panoramic images of the tunnel face. During acquisition, multiple shooting points and stitching ensure a wide recognition range and clear detail in the panoramic images. Each time an image is captured, a ranging instrument records the actual distance between the binocular camera and the tunnel face, which is then converted into the true width and height of each image frame, serving as the scaling reference before stitching. This allows the shooting process to be unrestricted by the harsh environmental conditions of underground engineering projects, enabling flexible shooting from any flat and safe location. During each shot, a compass or other orientation instrument is used to measure and calibrate the shooting direction, ensuring that the direction remains consistent along the tunnel axis for each shot. This guarantees the accuracy of the stitched panoramic image of the tunnel face, and the measured shooting direction also provides a basis for subsequent identification of fracture inclination information.

[0033] The working face image extraction device of the present invention is used to implement the above method. Its overall structure is simple and portable. The shooting position is not limited by the severe environmental conditions of underground engineering. The shooting point can be selected arbitrarily, reducing safety hazards. The distance parameters are directly obtained during shooting, making the stitched overall view of the working face more detailed, which is conducive to improving the accuracy of crack identification. In addition, the shooting direction can also be recorded to provide a basis for subsequent crack tendency identification. Attached Figure Description

[0034] Figure 1 This is a schematic diagram illustrating the principle of step S4 in the image analysis and calculation process of a method for identifying cracks on the face of a working face according to the present invention.

[0035] Figure 2 This is a schematic diagram of the image stitching process in step S5 of the method for identifying cracks on the working face of the present invention;

[0036] Figure 3 This is an example of vectorization operation of the crack trace image in step S6 of the method for identifying cracks on the face of the machine according to the present invention;

[0037] Figure 4 This is a schematic diagram of the overall structure of a face image extraction device according to the present invention;

[0038] Figure 5 for Figure 4 The diagram shows the structure after removing the tripod. Detailed Implementation

[0039] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the scope of protection of the present invention is not limited to the following description.

[0040] A method for identifying cracks at the tunnel face includes the following steps:

[0041] S1. Set up a central camera point and multiple peripheral camera points in front of the tunnel face. The distance between each camera point and the tunnel face is not limited; any flat, safe area in front of the tunnel face can be selected. The central camera point is set on the tunnel axis in front of the tunnel face. The peripheral camera points are set arbitrarily according to actual needs. The number of peripheral camera points and their lateral positions must ensure that the images of the tunnel face captured by all camera points can be stitched together to form a complete panoramic image of the tunnel face with a wide recognition range.

[0042] S2. Take pictures of the tunnel face from the central shooting point using a binocular camera. During the shooting, the binocular camera should be kept horizontal and facing the tunnel face. Record the shooting direction of the binocular camera using a compass or other orientation instrument. After taking the picture, import the captured image into the computer and record the actual distance between the binocular camera and the tunnel face when the picture was taken using a rangefinder. At the same time, record the focal length, magnification, and physical parameter information of the binocular camera when taking the picture (the parameters of the binocular camera itself, including the critical focal length of the optical zoom and digital zoom of the binocular camera, the critical magnification of the optical zoom and digital zoom of the binocular camera, the center line of the two lens sensors in the binocular camera, the height of the lens sensors in the binocular camera, the width of the lens sensors in the binocular camera, etc.). When shooting from the central shooting point, the binocular camera takes pictures at different heights in the vertical direction in the manner described above, so that the pictures contain all the information about the position of the tunnel face on the tunnel axis in the vertical direction. At the same time, for each shot, the actual distance between the binocular camera and the tunnel face, the focal length of the binocular camera, the magnification, and the physical parameter information of each picture are recorded in the manner described above.

[0043] S3. Using a binocular camera, take pictures at different heights from various surrounding shooting points. Keep the binocular camera horizontal during each shot, and adjust its shooting direction using a compass or other orientation instruments to ensure it remains consistent with the direction used when shooting from the central shooting point. Similarly, import the captured images into a computer and record the actual distance between the binocular camera and the working face, the camera's focal length, magnification, and physical parameters for each shot.

[0044] S4. The computer analyzes and calculates the imported data. Based on the actual distance between the stereo camera and the working face for each image, the focal length and magnification of the stereo camera, and the physical parameters of the stereo camera at each shooting time, it calculates the true width and true height of the image. For specific calculations, please refer to [link / reference needed]. Figure 1 The formulas for calculating the actual width and height of an image are as follows:

[0045] When N≤N0:

[0046] ,

[0047] ,

[0048] When N > N0:

[0049] ,

[0050] ,

[0051] in,

[0052] D is the actual width of the image.

[0053] H represents the actual height of the image.

[0054] f is the focal length of the stereo camera when taking the picture;

[0055] f0 is the critical focal length between optical zoom and digital zoom of a stereo camera, and is a physical parameter of the stereo camera.

[0056] N is the magnification of the stereo camera when taking pictures;

[0057] N0 is the critical magnification between optical zoom and digital zoom of the stereo camera, and is a physical parameter of the stereo camera.

[0058] L represents the actual distance between the binocular camera and the working face during the shooting process;

[0059] B is the center line of the two lens sensors in the stereo camera, which is a physical parameter of the stereo camera;

[0060] h represents the height of the lens sensor in the stereo camera, and represents the physical parameters of the stereo camera.

[0061] d represents the width of the lens sensor in the stereo camera, and d represents the physical parameters of the stereo camera.

[0062] S5. The computer uses an image stitching toolbox to scale the images taken at various heights from different shooting points, based on the actual width and height of the captured images. Then, it... Figure 2 As shown, based on the overlapping details in each scaled image, the scaled images are stitched together to form a complete panoramic image of the tunnel face, and the true width and true height corresponding to this panoramic image are calculated. For example, when stitching images A and B taken from similar shooting points, images A and B, taken directly by the same binocular camera, have the same size. Let the true width corresponding to the length of image A be 3m, and the true height corresponding to the width of image A be 2m. Let the true width corresponding to the length of image B be 4.5m, and the true height corresponding to the width of image B be 3m. Before stitching, image B can be enlarged by 1.5 times based on the true length or true width corresponding to image A before stitching images A and B.

[0063] Therefore, through steps S1 to S5, a panoramic image of the tunnel face can be acquired. During acquisition, detailed shots are taken from multiple shooting points and then stitched together, ensuring a wide recognition range and clear details in the panoramic image. During each shot, a rangefinder records the actual distance between the binocular camera and the tunnel face, which is then converted into the true width and height of each image frame as the scaling reference before stitching. This allows the shooting process to be unrestricted by the harsh environmental conditions of underground engineering, enabling flexible shooting from any flat and safe location. During each shot, a compass or other orientation instrument is used to measure and calibrate the shooting direction, ensuring that the direction remains consistent along the tunnel axis, guaranteeing the accuracy of the stitched panoramic image. Furthermore, since the binocular camera is kept horizontal and facing the tunnel face during shooting in step S2, the determined shooting direction is the tunnel axis direction, and this measured shooting direction also provides a basis for subsequent identification of fracture inclination information.

[0064] S6. The computer identifies and extracts crack trace information from the panoramic image of the tunnel face using the U-net network model, realizing the vectorization of the crack trace image. The operation process is as follows: Figure 3 As shown.

[0065] S7. The computer imports the vectorized image into the fracture orientation extraction system to obtain information on fracture length, fracture dip angle, and fracture dip direction. Specifically:

[0066] Obtaining fracture length information involves importing the vectorized image into a fracture orientation extraction system, extracting the coordinates of each pixel in the image, using the 8-neighborhood tracking method to find fracture endpoints and trace the fracture path, thereby extracting the skeleton of the entire fracture; then, calculating the distance between consecutive points on the fracture path using the Euclidean distance formula, and summing them to obtain the length of the entire fracture; finally, converting the pixel length of the fracture into its actual length based on the true width and height corresponding to the panoramic image of the tunnel face; the calculation formula is as follows:

[0067] ,

[0068] Where l is the actual length of the crack, l i Let be the pixel distance between two adjacent points on the crack, D be the actual length of the image, and b be the pixel length of the image. i y i () represents the coordinates of a pixel in the crack skeleton.

[0069] Obtaining fracture dip angle information involves importing the vectorized image into a fracture orientation extraction system, extracting the coordinates of each pixel in the image, and then performing a linear regression algorithm combined with a loop function to linearly fit all pixel coordinates of a single fracture skeleton to obtain the slope of the line. The slope is then converted into an angle to obtain the apparent dip angle of the fracture. The specific calculation formula is as follows:

[0070] ,

[0071] Where k is the slope of the line, and c is the y-intercept of the line. The apparent dip angle of the fracture. , , (x i y i () represents the coordinates of a pixel in the crack skeleton.

[0072] Obtaining the fracture dip direction refers to calculating the fracture dip direction based on the shooting direction of the binocular camera (i.e., the direction of the tunnel axis) and the apparent dip angle of the fracture. The calculation formula is as follows:

[0073] ,

[0074] in, The direction of the tunnel axis, The azimuth angle of the face is . The angle of the fracture dip. The calculated fracture tendency; The fracture tends to If the calculated The fracture tends to .

[0075] like Figure 4 , Figure 5As shown, a face image extraction device includes a computer and a binocular camera 1. The binocular camera 1 is equipped with an infrared range sensor 2 and an electronic compass 3. The infrared range sensor 2 is installed at the front end of the binocular camera 1 and is located at the midpoint of the center line between the two lens sensors of the binocular camera 1. The infrared range sensor 2 is used to detect the distance between the obstacle to be photographed in front of the binocular camera 1 and the binocular camera 1. The electronic compass 3 is used to detect the shooting direction of the binocular camera 1. The binocular camera 1, the infrared sensor 2, and the electronic compass 3 are all electrically connected to the computer. This face image extraction device is used to implement the aforementioned face crack identification method. The electronic compass 3 detects and calibrates the shooting direction of the binocular camera 1 during each shot, while the infrared distance sensor 2 measures the actual distance between the binocular camera 1 and the face during each shot. A portable computer can be used to receive and compare the shooting direction detected by the electronic compass 3 during each shot, guiding operators to ensure consistent shooting directions. The computer also receives images captured each time, relevant parameters of the binocular camera 1 during each shot, and distance values ​​detected by the infrared sensor 2 during each shot, and completes the analysis, calculation, image stitching, and crack identification processes described in S4 to S7. Overall, this face image extraction device has a simple and portable structure. The shooting location is not limited by the harsh environmental conditions of underground engineering, and the shooting point can be arbitrarily selected, reducing safety hazards. Directly acquiring distance parameters during shooting makes the stitched face image more detailed, improving the accuracy of crack identification. Furthermore, it can record the shooting direction, providing a basis for subsequent crack identification.

[0076] Furthermore, it also includes a leveling base 4, whose structure is similar to that of the leveling platform of engineering theodolites, levels, etc. It includes a base plate 5 and a top plate 6, which are connected by three leveling screws 7. A level 8 is installed on the top plate 6. After the base plate 5 is fixed to the bracket, the top plate 6 can be adjusted to a horizontal state using the leveling screws 7. The level 8 can be a bubble level, used to observe the leveling state during leveling. A binocular camera 1 is set on the top plate 6, and the shooting direction of the binocular camera 1 is parallel to the top plate 6. In use, the base plate 5 can be mounted on any bracket for support. The leveling base 4 levels the top plate 6 so that the binocular camera 1 can accurately shoot in the horizontal direction, maintaining the consistency of the shooting direction at each shooting point. In specific implementation, it also includes a tripod 9 commonly used in engineering inspection, with the base plate 5 directly and detachably connected to the tripod head of the tripod 9.

[0077] Furthermore, it also includes a rotating base 10, which is connected to the top plate 6. The binocular camera 1 is rotatably connected to the rotating base 10. By rotating, the shooting direction of the binocular camera 1 can be adjusted. It can be rotated at the central shooting point to make the binocular camera 1 face the working face directly, or after the tripod is placed and leveled at each of the surrounding shooting points, the shooting direction of the binocular camera 1 can be aligned at each shooting point by rotating. A locking screw is provided on the rotating base 10, which can be used to lock the binocular camera 1 after adjustment.

[0078] Furthermore, it also includes a screw jack 11, a commercially available device. The screw jack 11 drives a worm gear to rotate, and the worm gear and screw form a screw-nut pair. When the worm gear rotates, it can drive the screw to rise and fall. The housing of the screw jack 11 is fixedly connected to the top plate 6, and the screw of the screw jack 11 is perpendicular to the top plate 6. The rotary seat 10 is fixedly installed on the top of the screw of the screw jack 11. Rotating the worm gear of the screw jack 11 can adjust the height of the rotary seat 10 and the binocular camera 1, so as to obtain images of the tunnel face at different heights at each shooting point, facilitating the formation of a complete overall image of the tunnel face after stitching.

[0079] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A method for identifying cracks at the working face, characterized in that, Includes the following steps: S1. A central shooting point and multiple peripheral shooting points are set in front of the working face. The central shooting point and peripheral shooting points are set in flat and safe areas according to the actual situation. The central shooting point is set on the tunnel axis in front of the working face. S2. Take pictures at different heights at the central shooting point using a binocular camera. During shooting, the binocular camera is kept horizontal and facing the working face. Record the shooting direction of the binocular camera during shooting. Import each picture into the computer and record the actual distance between the binocular camera and the working face, the focal length of the binocular camera and the magnification of the binocular camera during each shooting. S3. Take pictures at different heights at each of the surrounding shooting points using a binocular camera. During shooting, the binocular camera is kept horizontal and the shooting direction is consistent with the direction when shooting at the central shooting point. Import the captured pictures into the computer and record the actual distance between the binocular camera and the working face, the focal length of the binocular camera and the magnification of the binocular camera at each shooting. S4. The computer analyzes and calculates the imported data, and calculates the actual width and height of the image based on the actual distance between the binocular camera and the working face corresponding to each image, the focal length and magnification of the binocular camera, and the physical parameters of the binocular camera. S5. The computer uses an image stitching toolbox to scale the image according to the actual width and actual height of the captured image, and then completes the stitching of the panorama image of the working face to obtain the panorama image of the working face and calculates the actual width and actual height of the panorama image of the working face. S6. The computer identifies and extracts crack trace information in the panoramic image of the tunnel face through the U-net network model, thereby realizing the vectorization of the crack trace image; S7. The computer imports the vectorized image into the fracture orientation extraction system to obtain fracture length, fracture dip angle, and fracture dip information. The fracture length information refers to the process of importing the vectorized image into a fracture orientation extraction system, extracting the coordinates of each pixel in the image, using the 8-neighborhood tracking method to find the fracture endpoints and tracing the fracture path to extract the skeleton of the entire fracture, then calculating the distance between consecutive points on the fracture path according to the Euclidean distance formula, accumulating these distances to obtain the pixel length of the entire fracture, and finally converting the pixel length of the fracture into the actual length based on the true width and true height corresponding to the panoramic image of the tunnel face. The fracture dip angle information refers to importing the vectorized image into a fracture orientation extraction system, extracting the coordinates of each pixel in the image, and then performing a linear regression algorithm combined with a loop function to linearly fit all pixel coordinates of a single fracture skeleton to obtain the slope of the straight line. The slope is then converted into an angle to obtain the apparent dip angle of the fracture. The fracture tendency information refers to the fracture tendency calculated based on the shooting direction of the binocular camera and the fracture viewing angle.

2. The method for identifying face cracks according to claim 1, characterized in that, The formulas for calculating the actual width and actual height of the image in step S4 are as follows: When N≤N0: ， ， When N > N0: ， ， Wherein, D is the actual width of the image; H is the actual height of the image; f is the focal length of the stereo camera when taking the image; f0 is the critical focal length between optical zoom and digital zoom of the stereo camera, which is a physical parameter of the stereo camera; N is the magnification of the stereo camera when taking the image; N0 is the critical magnification between optical zoom and digital zoom of the stereo camera, which is a physical parameter of the stereo camera; L is the actual distance between the stereo camera and the working face when taking the image; B is the center line of the two lens sensors in the stereo camera, which is a physical parameter of the stereo camera; h is the height of the lens sensors in the stereo camera, which is a physical parameter of the stereo camera; and d is the width of the lens sensors in the stereo camera, which is a physical parameter of the stereo camera.

3. A device for extracting images of a working face, characterized in that, A method for identifying face cracks according to claim 1 includes a computer and a binocular camera. The binocular camera is equipped with an infrared range sensor and an electronic compass. The infrared range sensor is installed at the front end of the binocular camera and is positioned at the midpoint of the center line between the two lens sensors of the binocular camera. The infrared range sensor is used to detect the distance between the obstacle being photographed in front of the binocular camera and the binocular camera. The electronic compass is used to detect the shooting direction of the binocular camera. The binocular camera, infrared sensor, and electronic compass are all electrically connected to the computer.

4. The face image extraction device according to claim 3, characterized in that, It also includes a leveling base, which includes a base plate and a top plate connected by three anchor bolts. A level is installed on the top plate, and the binocular camera is mounted on the top plate with its shooting direction parallel to the top plate.

5. The face image extraction device according to claim 4, characterized in that, It also includes a rotary base, which is connected to the top plate, and the binocular camera is rotatably connected to the rotary base.

6. The face image extraction device according to claim 5, characterized in that, It also includes a screw jack, the housing of which is fixedly connected to the top plate, the screw of which is perpendicular to the top plate, and the rotary seat is fixedly installed on the top of the screw of which is the screw jack.

7. The face image extraction device according to claim 4, characterized in that, It also includes a tripod, the tripod head of which is detachably connected to the base plate.

Images