Transparent soil visualization test equipment and image processing method

Abstract

The invention belongs to the technical field of cast-in-situ bored pile post-grouting engineering, and particularly relates to transparent soil visualization test equipment and an image processing method.The transparent soil visualization test equipment comprises a supporting frame, an angle adjusting mechanism is arranged at the upper end of the supporting frame, bases are symmetrically arranged on the two sides of the supporting frame, and servo motors are embedded in the middles of the bases; and an output shaft of the servo motor is fixedly connected with a U-shaped rotating rod. According to the visual test equipment, the industrial camera, the high-resolution CCD, the polarization filter, the focusable lens and the damping module are arranged, so that mud reflection and vibration interference are eliminated, the image fidelity and the focusing success rate are greatly improved, micron-sized detail data are provided, and the visual test efficiency is improved. And meanwhile, by combining algorithm dynamic refractive index correction, self-adaptive preprocessing, three-dimensional displacement field calculation and quantitative analysis, displacement errors are effectively reduced, full-dimensional deformation reduction of the soil body is realized, and a strain and disturbance radius quantitative report is output.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of post-grouting engineering technology of cast-in-place pile, and in particular to a transparent soil visual test device and an image processing method. BACKGROUND

[0002] The post-grouting technology of cast-in-place pile is a core technology for foundation reinforcement, bearing capacity and anti-settlement capacity of pile foundation in construction engineering. The accurate revelation of the grouting diffusion path and the micro disturbance law of the soil body requires transparent soil visual test. The transparent soil simulates the physical properties of the real soil body, and the deformation response of the soil body during drilling and grouting is analyzed intuitively through optical observation, thereby providing mechanism support for engineering design.

[0003] However, the existing transparent soil visual test still has some problems, such as: 1. Single drilling angle. Most devices can only achieve vertical drilling, and cannot simulate the inclined drilling of slope pile foundation and the inclined drilling under complex geology in actual engineering, resulting in a disconnection between the test scene and the actual engineering, and limited data reference value; 2. Low image fidelity. Industrial cameras are easily disturbed by mud particle reflection, and the vibration of servo motors can cause image blur. The low-resolution sensor and fixed-focus lens are difficult to capture the micron-level displacement of transparent soil particles, and subsequent data processing is prone to errors.

[0004] In view of the above problems, the present application provides a transparent soil visual test device and an image processing method. SUMMARY

[0005] In view of the technical problems of single drilling angle, low image fidelity of industrial cameras, and low automation degree of the test, the present application provides a transparent soil visual test device and an image processing method.

[0006] The transparent soil visual test device provided by the present application comprises a support frame, an angle adjusting mechanism is arranged at the upper end of the support frame, two bases are symmetrically arranged at the two sides of the support frame, a servo motor is embedded in the middle position of the base, a U-shaped rotating rod is fixedly connected to the output shaft of the servo motor, industrial cameras are arranged in an array on the inner side of the rotating rod, an observation basin is fixedly connected to the inner side wall of the upper end of the support frame, the bottom and the side of the observation basin are rotatably connected through a pin shaft, and a gas cylinder is symmetrically arranged on the lower end of the observation basin along the width direction axis. The angle adjusting mechanism is used to adjust the different angle drilling of the transparent soil in the observation basin.

[0007] Preferably, the angle adjustment mechanism includes an arched slide, the body of which has a cross groove, a mounting frame slidably connected to the surface of the slide, and the mounting frame having a large and two small hollow slots inside. A dual-head motor is fixedly connected to the inner wall of the large hollow slot of the mounting frame, and a main gear is rotatably connected to the inner wall of the small hollow slot of the mounting frame via bearings. The two output shafts of the dual-head motor are fixedly connected to the sides of the main gear, and a hydraulic cylinder is fixedly connected to the side of the mounting frame. A drill bit is fixedly connected to the piston rod end of the hydraulic cylinder. Both ends of the mounting frame are slidably connected to the inner wall of the cross groove, and connecting rods are fixedly connected to both ends of the mounting frame. The connecting rods are located inside the cross groove, and one end of the connecting rod is fixedly connected to the surface of the hydraulic cylinder.

[0008] Preferably, the outer surface array of the carriage is configured as a toothed surface adapted to the surface of the main gear.

[0009] Preferably, the industrial camera has a built-in high-resolution CCD sensor and an adjustable focus lens with a polarizing filter embedded in the front end, and a controller is also provided on one side of the industrial camera.

[0010] Preferably, the maximum opening angle of the lower end of the observation basin is a horizontal angle of 45 degrees.

[0011] Preferably, the controller integrates a real-time mud three-parameter detection module, and a specific gravity sensor, a viscosity sensor, and a sand content sensor are embedded in the inner wall of the observation basin. The controller has a built-in dynamic calculation model for mud refractive index, which corrects the equivalent refractive index of transparent soil in real time based on the detected mud parameters. The formula is as follows: ,in, The refractive index of the transparent clay substrate is taken as 1.45~1.46. The sand content of the mud. The refractive index of the mud base is taken as 1.43~1.44. 0.3 represents the mud viscosity and 0.3 represents the mud proportion weight.

[0012] Preferably, S1, angle adjustment The mechanism's dual-head motor drives the main gear to move along the toothed surface of the arched slide, moving the mounting frame, connected hydraulic cylinder, and drill bit to the initial drilling position. The drilling angle is [angle missing]. , The preset angles are 30°, 60° and 90°. The controller controls the hydraulic cylinder to drive the drill bit to start drilling. At the same time, the industrial camera is triggered to start collecting images. The frame rate is linked to the drilling speed of the drill bit. The servo motor starts to drive the U-shaped rotating rod to rotate, which in turn drives the industrial camera to take pictures of the transparent soil being drilled from multiple angles. Meanwhile, the vibration reduction module built into the base reduces the interference of servo motor vibration on the camera, ensuring image clarity. S2. After the drill bit completes drilling, the cylinder piston rod retracts, opening the lower end of the observation basin while simultaneously vibrating at a high frequency until the lower end of the observation basin reaches a maximum opening angle of 45 degrees. After the transparent soil inside the observation basin is discharged, the cylinder drives the lower end of the observation basin to close again. The operator adds another prepared portion of transparent soil into the observation basin through external equipment. After the operator completes the operation, the servo motor takes multiple photos from different angles again, and the industrial camera transmits the captured data to the cloud for processing.

[0013] Preferably, the multi-angle images acquired by the industrial camera in step S2 need to be preprocessed, specifically including adaptive denoising, layered refraction correction, and multi-view image registration. The preprocessed images are used for soil displacement field calculation, and the adaptive denoising is based on the mud sand content. The filter window is dynamically adjusted using the following formula: W is the side length of the filter window. The function is a floor function with a window range of 3×3 to 7×7 pixels. It eliminates the reflection and vibration interference of mud particles. In image processing, a 1×1 window has no denoising effect, a 2×2 window is prone to causing image edge shift, and a 3×3 window is the smallest odd-numbered window that balances denoising ability and detail preservation. It can filter out isolated noise of individual pixels without blurring the edge features of transparent soil particles. The layered refraction correction is based on Snell's law to correct image pixel coordinates, and the formula is as follows: ,in All are original pixel coordinates. This represents the refractive offset, and 1.45 is the standard refractive index reference value for transparent soil. The multi-view image registration is based on the rotation angle of the servo motor. Based on this, establish a perspective transformation matrix. The image from different angles is unified to the world coordinate system, with a registration error of ≤0.5 pixels.

[0014] Preferably, in step S2, the preprocessed image is used to derive the three-dimensional displacement field of the soil using an improved incremental RG-PIV algorithm, and a corner detection algorithm is employed. Extract feature points of transparent soil particles, among which These are the eigenvalues ​​of the image grayscale matrix; The multi-view image registration uses Normalized Cross-Correlation Coefficient (NCC) feature point matching, combined with mud viscosity. Calculate plane displacement ; Combined with the industrial camera, different preset angles are used based on the horizontal angle with the transparent soil. Establish a multi-view triangulation model to measure planar displacement. Transformation into three-dimensional displacement in world coordinate system The formula is ; ; L is the baseline distance between the two cameras, and f is the focal length of the industrial camera. Planar displacement from the perspective of view Planar displacement from the perspective of view Planar displacement from the perspective of view This is the refractive index correction factor.

[0015] Preferably, the soil strain field and borehole disturbance quantitative indicators are further derived based on the three-dimensional displacement field. The specific formulas and steps are as follows: P1. Calculate the spatial gradient of the smoothed three-dimensional displacement field to obtain the normal strain. With shear strain , , ,in These correspond to the horizontal radial deformation, the radial deformation of the borehole, the horizontal circumferential deformation, the circumferential deformation of the borehole, and the axial deformation of the borehole in the vertical depth direction, respectively. This represents the change in displacement for every 1 mm change in spatial position along the x-direction. These represent the shear strain in the xy plane and the shear strain in the xz plane, respectively. P2, Radius affected by borehole disturbance , The spatial coordinates of the drill bit center. The coordinates of feature points whose displacement exceeds a threshold value of 0.1 mm. In pixel equivalents; P3. The controller automatically generates a three-dimensional displacement field cloud map, strain distribution curve, and disturbance influence radius data, and associates them with the drill bit drilling angle and mud three parameters to form a quantitative test report.

[0016] The beneficial effects of this invention are as follows: 1. By setting an angle adjustment mechanism, the drill bit can be accurately switched between preset drilling angles of 30°, 60°, and 90° with an angle error of ≤0.5°. This solves the limitation of traditional equipment that can only drill at a single angle. This setting can simulate actual drilling conditions such as inclined and vertical drilling. Moreover, the connecting rod can stabilize the posture of the hydraulic cylinder and ensure that there is no deviation during drilling. It provides a controllable variable for analyzing the impact of drilling at different angles on soil disturbance, and improves the pertinence and reliability of the test data.

[0017] 2. By setting up an industrial camera with a high-resolution CCD, polarizing filter, adjustable focus lens and vibration reduction module, the interference of mud reflection and vibration is eliminated, which greatly improves the image fidelity and focusing success rate, and provides micron-level detail data. At the same time, combined with algorithm dynamic refractive index correction, adaptive preprocessing, three-dimensional displacement field calculation and quantitative analysis, the displacement error is effectively reduced, the full-dimensional deformation of the soil is restored, and the strain and disturbance radius quantitative report is output. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of a transparent soil visualization testing device proposed in this invention; Figure 2 This invention proposes a transparent soil visualization testing device. Figure 1 Enlarged view in the middle; Figure 3 This is a front view of a transparent soil visualization testing device proposed in this invention; Figure 4 This is a cylinder position diagram of a transparent soil visualization testing device proposed in this invention; Figure 5 This invention proposes a transparent soil visualization testing device. Figure 4 Enlarged view of point B in the middle; Figure 6 This is a flowchart of a transparent soil visualization test device and image processing method proposed in this invention.

[0019] In the diagram: 1. Support frame; 2. Angle adjustment mechanism; 20. Cross slide; 21. Carriage; 22. Mounting frame; 23. Dual-head motor; 24. Main gear; 25. Hydraulic cylinder; 26. Drill bit; 27. Connecting rod; 3. Base; 4. Servo motor; 5. Rotating rod; 6. Industrial camera; 7. Observation basin; 8. Cylinder. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0021] Reference Figures 1-6 A transparent soil visualization test device includes a support frame 1, an angle adjustment mechanism 2 at the upper end of the support frame 1, bases 3 symmetrically arranged on both sides of the support frame 1, a servo motor 4 embedded in the middle of the base 3, a U-shaped rotating rod 5 fixedly connected to the output shaft of the servo motor 4, an industrial camera 6 arrayed on the inner side of the rotating rod 5, an observation basin 7 fixedly connected to the inner wall of the upper end of the support frame 1, the bottom of the observation basin 7 and its side are rotatably connected by a pin, and a cylinder 8 is symmetrically arranged along the width direction at the lower end of the observation basin 7.

[0022] The angle adjustment mechanism 2 is designed to adjust the drilling angle of the transparent soil in the observation basin 7.

[0023] In this embodiment, the angle adjustment mechanism 2 includes an arched slide 21. The body of the slide 21 has a cross groove 20. A mounting frame 22 is slidably connected to the surface of the slide 21. The mounting frame 22 has a large and two small hollow slots inside. A dual-head motor 23 is fixedly connected to the inner wall of the large hollow slot of the mounting frame 22. A main gear 24 is rotatably connected to the inner wall of the small hollow slot of the mounting frame 22 through a bearing. The two output shafts of the dual-head motor 23 are fixedly connected to the sides of the main gear 24. A hydraulic cylinder 25 is fixedly connected to the side of the mounting frame 22. A drill bit 26 is fixedly connected to the piston rod end of the hydraulic cylinder 25. The two ends of the mounting frame 22 are slidably connected to the inner wall of the cross groove 20. A connecting rod 27 is fixedly connected to the two ends of the mounting frame 22. The connecting rod 27 is located inside the cross groove 20. One end of the connecting rod 27 is fixedly connected to the surface of the hydraulic cylinder 25. The outer surface of the slide 21 is arrayed with toothed surfaces that are adapted to the surface of the main gear 24.

[0024] Specifically, the toothed surfaces on the outer surface of the slide 21 are arranged in an equidistant array, with the tooth pitch and height matching the tooth profile parameters of the main gear 24. There is no tooth backlash, avoiding problems such as jamming or offset during angle adjustment. This improves the drilling angle control accuracy of the drill bit 26 to ±0.2°, ensuring the consistency of multiple sets of parallel test data. The main gear 24 meshes with the toothed surfaces on the outer surface of the slide 21. When the dual-head motor 23 starts, its output shaft drives the main gear 24 to roll along the toothed surfaces, thereby driving the mounting frame 22 to slide smoothly and linearly along the cross groove 20. Simultaneously, the connecting rods 27 at both ends of the mounting frame 22 synchronously pull the hydraulic cylinder 25, ensuring that the hydraulic cylinder 25 and the drill bit 26 maintain a preset posture during sliding. Ultimately, this achieves precise switching of the drill bit 26 between preset drilling angles of 30°, 60°, and 90°, with an angle adjustment error ≤0.5°, meeting the requirements of transparent soil drilling tests under different working conditions.

[0025] In this embodiment, the industrial camera 6 has a built-in high-resolution CCD sensor and an adjustable focus lens with a polarizing filter embedded in the front end. A controller is also provided on one side of the industrial camera 6.

[0026] Specifically, the high-resolution CCD sensor of the industrial camera 6 has a pixel value of ≥5 million, which can clearly capture the micron-level displacement characteristics of transparent soil particles; the polarizing filter at the front end can filter the diffuse reflection light of suspended particles in the mud, reducing image noise interference; the focal length of the adjustable lens has a range of 10-50mm, which can be adjusted in real time according to the drilling depth of the drill bit 26 to ensure clear image focus; the controller on one side links the acquisition frame rate of the industrial camera 6 to match the drilling speed of the drill bit in a 1:1 ratio, with a speed range of 0.5-5mm / s, and receives data from specific gravity, viscosity, and sand content sensors in real time, and simultaneously performs mud refractive index correction and image preprocessing.

[0027] In this embodiment, the maximum opening angle of the lower end of the observation basin 7 is a horizontal angle of 45 degrees.

[0028] Specifically, by observing the maximum opening angle of the bottom of the basin 7 at 45 degrees, and by observing the high-frequency vibration of the basin 7, the transparent soil after the test can be quickly discharged under the combined action of gravity and vibration, leaving no residue and no adhesion to the basin wall. This facilitates the quick replacement of the transparent soil for the next test and analysis at different angles.

[0029] In this embodiment, the controller integrates a real-time mud three-parameter detection module, observing the specific gravity sensor, viscosity sensor, and sand content sensor embedded in the inner wall of the observation basin 7. The controller has a built-in dynamic calculation model for the mud refractive index, which corrects the equivalent refractive index of the transparent soil in real time based on the detected mud parameters. The formula is as follows: ,in, The refractive index of the transparent clay substrate is taken as 1.45~1.46. The sand content of the mud. The refractive index of the mud base is taken as 1.43~1.44. 0.3 represents the mud viscosity and 0.3 represents the mud proportion weight.

[0030] Specifically, the specific gravity sensor, viscosity sensor, and sand content sensor on the inner wall of basin 7 are observed to collect mud parameters in real time and transmit them to the controller. The controller's built-in dynamic refractive index calculation model converts the sensor data into an equivalent refractive index correction value through the above formula. This can offset the influence of mud sand content and viscosity changes on light propagation, avoid image pixel shift caused by refractive index deviation, and provide accurate basic parameters for subsequent layered refractive correction.

[0031] refer to Figures 1-6 An image processing method for a transparent soil visualization testing device, the specific implementation steps of which are as follows: S1. The dual-head motor 23 of the angle adjustment mechanism 2 drives the main gear 24 to move along the tooth surface of the arched slide 21, thereby moving the mounting frame 22 and the connected hydraulic cylinder 25 and drill bit 26 to the initial drilling position. The drilling angle is the same as the horizontal angle. , The preset angles are 30°, 60° and 90°. The controller controls the hydraulic cylinder 25 to drive the drill bit 26 to start drilling. At the same time, the industrial camera 6 is triggered to start collecting images. The frame rate is linked to the drilling speed of the drill bit 26. The servo motor 4 starts to drive the U-shaped rotating rod 5 to rotate, which in turn drives the industrial camera 6 to take pictures of the drilled transparent soil from multiple angles. At the same time, the shock absorption module built into the base 3 reduces the interference of the vibration of the servo motor 4 on the camera, ensuring image clarity. After drilling S2 and drill bit 26 completes drilling, the piston rod of cylinder 8 retracts, opening the lower end of observation basin 7 while vibrating at high frequency until the lower end of observation basin 7 reaches a maximum opening angle of 45 degrees. After the transparent soil inside observation basin 7 is discharged, cylinder 8 drives the lower end of observation basin 7 to close again. The operator adds another prepared portion of transparent soil into observation basin 7 through external equipment. After the operator completes the operation, servo motor 4 takes multiple photos again, and industrial camera 6 transmits the captured data to the cloud for processing.

[0032] In this embodiment, the multi-angle images acquired by the industrial camera 6 in S2 need to be preprocessed, specifically including adaptive denoising, layered refraction correction, and multi-view image registration. The preprocessed images are used for soil displacement field calculation, and the adaptive denoising is based on the mud sand content. The filter window is dynamically adjusted using the following formula: W is the side length of the filter window. The function is a floor function with a window range of 3×3 to 7×7 pixels. It eliminates the reflection and vibration interference of mud particles. In image processing, a 1×1 window has no denoising effect, a 2×2 window is prone to causing image edge shift, and a 3×3 window is the smallest odd-numbered window that balances denoising ability and detail preservation. It can filter out isolated noise of individual pixels without blurring the edge features of transparent soil particles. Layered refraction correction corrects image pixel coordinates based on Snell's law, using the following formula: ,in All are original pixel coordinates. This represents the refractive offset, and 1.45 is the standard refractive index reference value for transparent soil. Multi-view image registration based on the rotation angle of servo motor 4 Based on this, establish a perspective transformation matrix. The image from different angles is unified to the world coordinate system, with a registration error of ≤0.5 pixels.

[0033] Specifically, in the adaptive noise reduction process, when the mud sand content... When =0.1, the filter window =5, that is, 5×5 pixels, can filter out medium-sized grainy reflections; when When W=0.2, W=7, which is 7×7 pixels, it is suitable for noisy scenes with high sand content mud, and always retains the edge features of transparent soil particles. The multi-view image matching criterion is based on the rotation angle of servo motor 4. Based on this, establish a perspective transformation matrix. Images from 30°, 60°, and 90° perspectives are unified to the world coordinate system, with registration errors strictly controlled to ≤0.5 pixels, ensuring compatibility of multi-view data.

[0034] In this embodiment, the preprocessed image in S2 is used to derive the three-dimensional displacement field of the soil using the improved incremental RG-PIV algorithm, and a corner detection algorithm is employed. Extract feature points of transparent soil particles, among which These are the eigenvalues ​​of the image grayscale matrix; For multi-view image registration, normalized cross-correlation coefficient (NCC) is used to match feature points, combined with mud viscosity. Calculate plane displacement ; Combined with an industrial camera 6, different preset angles are used based on the horizontal angle relative to the transparent soil. Establish a multi-view triangulation model to measure planar displacement. Transformation into three-dimensional displacement in world coordinate system The formula is ; ; L is the baseline distance between the two cameras, and f is the focal length of the industrial camera. Planar displacement from the perspective of view Planar displacement from the perspective of view Planar displacement from the perspective of view This is the refractive index correction factor.

[0035] In this embodiment, the soil strain field and borehole disturbance quantitative indicators are further derived based on the three-dimensional displacement field. The specific formulas and steps are as follows: P1. Calculate the spatial gradient of the smoothed three-dimensional displacement field to obtain the normal strain. With shear strain , , ,in These correspond to the horizontal radial deformation, the radial deformation of the borehole, the horizontal circumferential deformation, the circumferential deformation of the borehole, and the axial deformation of the borehole in the vertical depth direction, respectively. This represents the change in displacement for every 1 mm change in spatial position along the x-direction. These represent the shear strain in the xy plane and the shear strain in the xz plane, respectively. P2, Radius affected by borehole disturbance , The spatial coordinates of the drill bit center. The coordinates of feature points whose displacement exceeds a threshold value of 0.1 mm. In pixel equivalents; P3. The controller automatically generates a three-dimensional displacement field cloud map, strain distribution curve, and disturbance influence radius data, and associates them with the drill bit drilling angle and mud three parameters to form a quantitative test report.

[0036] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A transparent soil visualization testing device, comprising a support frame (1), characterized in that: An angle adjustment mechanism (2) is provided at the upper end of the support frame (1). Bases (3) are symmetrically arranged on both sides of the support frame (1). A servo motor (4) is embedded in the middle of the base (3). A U-shaped rotating rod (5) is fixedly connected to the output shaft of the servo motor (4). An industrial camera (6) is arranged in an array on the inner side of the rotating rod (5). An observation basin (7) is fixedly connected to the inner wall of the upper end of the support frame (1). The bottom of the observation basin (7) is rotatably connected to its side by a pin. A cylinder (8) is symmetrically arranged along the width direction at the lower end of the observation basin (7). The angle adjustment mechanism (2) is used to adjust the drilling angle of the transparent soil in the observation basin (7) at different angles.

2. The transparent soil visualization testing device according to claim 1, characterized in that: The angle adjustment mechanism (2) includes an arched slide (21). The body of the slide (21) has a cross groove (20). A mounting frame (22) is slidably connected to the surface of the slide (21). The mounting frame (22) has one large and two small hollow slots inside. A dual-head motor (23) is fixedly connected to the inner wall of the large hollow slot of the mounting frame (22). A main gear (24) is rotatably connected to the inner wall of the small hollow slot of the mounting frame (22) through a bearing. The two output shafts of the dual-head motor (23) are respectively connected to the main... The gear (24) is fixedly connected to the side, and the mounting frame (22) is fixedly connected to the side with a hydraulic cylinder (25). The piston rod end of the hydraulic cylinder (25) is fixedly connected to a drill bit (26). The two ends of the mounting frame (22) are slidably connected to the inner wall of the cross groove (20). The two ends of the mounting frame (22) are fixedly connected to a connecting rod (27). The connecting rod (27) is located inside the cross groove (20). One end of the connecting rod (27) is fixedly connected to the surface of the hydraulic cylinder (25).

3. The transparent soil visualization testing device according to claim 2, characterized in that: The outer surface array of the carriage (21) is configured as a toothed surface adapted to the surface of the main gear (24).

4. The transparent soil visualization testing device according to claim 3, characterized in that: The industrial camera (6) has a built-in high-resolution CCD sensor and an adjustable focus lens with a polarizing filter embedded in the front end. A controller is also provided on one side of the industrial camera (6).

5. The transparent soil visualization testing device according to claim 4, characterized in that: The maximum opening angle of the lower end of the observation basin (7) is 45 degrees horizontally.

6. The transparent soil visualization testing device according to claim 5, characterized in that: The controller integrates a real-time detection module for three mud parameters. A specific gravity sensor, a viscosity sensor, and a sand content sensor are embedded in the inner wall of the observation basin (7). The controller has a built-in dynamic calculation model for the mud refractive index, which corrects the equivalent refractive index of transparent soil in real time based on the detected mud parameters. The formula is as follows: ,in, The refractive index of the transparent clay substrate is taken as 1.45~1.

46. The sand content of the mud. The refractive index of the mud base is taken as 1.43~1.

44. 0.3 represents the mud viscosity and 0.3 represents the mud proportion weight.

7. The image processing method for a transparent soil visualization testing device according to claim 6, characterized in that, The specific implementation steps are as follows: S1. The dual-head motor (23) of the angle adjustment mechanism (2) drives the main gear (24) to move along the tooth surface of the arched slide (21), thereby moving the mounting frame (22) and the connected hydraulic cylinder (25) and drill bit (26) to the initial drilling position. The drilling angle is the same as the horizontal angle. , The preset angles are 30°, 60° and 90°. The controller controls the hydraulic cylinder (25) to drive the drill bit (26) to start drilling. Simultaneously, the industrial camera (6) is triggered to start collecting images. The frame rate is linked to the drilling speed of the drill bit (26). The servo motor (4) starts to drive the U-shaped rotating rod (5) to rotate, thereby driving the industrial camera (6) to take pictures of the drilled transparent soil from multiple angles. At the same time, the shock absorption module built into the base (3) reduces the interference of the servo motor (4) vibration on the camera, ensuring image clarity. S2. After the drill bit (26) finishes drilling, the piston rod of the cylinder (8) retracts, opening the lower end of the observation basin (7) and vibrating at high frequency until the lower end of the observation basin (7) reaches a maximum opening angle of 45 degrees. After the transparent soil inside the observation basin (7) is discharged, the cylinder (8) drives the lower end of the observation basin (7) to close again. The operator adds another prepared transparent soil into the observation basin (7) through an external device. After the operator completes the operation, the servo motor (4) takes pictures from multiple angles again, and the industrial camera (6) transmits the captured data to the cloud for processing.

8. The image processing method for a transparent soil visualization testing device according to claim 7, characterized in that, The multi-angle images acquired by the industrial camera (6) in S2 need to be preprocessed, specifically including adaptive denoising, layered refraction correction, and multi-view image registration. The preprocessed images are used for soil displacement field calculation. The adaptive denoising is based on the mud sand content. The filter window is dynamically adjusted using the following formula: W is the side length of the filter window. The function is a floor function with a window range of 3×3 to 7×7 pixels. It eliminates the reflection and vibration interference of mud particles. In image processing, a 1×1 window has no denoising effect, a 2×2 window is prone to causing image edge shift, and a 3×3 window is the smallest odd-numbered window that balances denoising ability and detail preservation. It can filter out isolated noise of individual pixels without blurring the edge features of transparent soil particles. The layered refraction correction is based on Snell's law to correct image pixel coordinates, and the formula is as follows: ,in All are original pixel coordinates. This represents the refractive offset, and 1.45 is the standard refractive index reference value for transparent soil. The multi-view image registration is based on the rotation angle of the servo motor (4). Based on this, establish a perspective transformation matrix. The image from different angles is unified to the world coordinate system, with a registration error of ≤0.5 pixels.

9. The image processing method for a transparent soil visualization test device according to claim 8, characterized in that, The preprocessed image in S2 uses an improved incremental RG-PIV algorithm to derive the three-dimensional displacement field of the soil, and employs a corner detection algorithm. Extract feature points of transparent soil particles, among which These are the eigenvalues ​​of the image grayscale matrix; The multi-view image registration uses Normalized Cross-Correlation Coefficient (NCC) feature point matching, combined with mud viscosity. Calculate plane displacement ; Combined with the industrial camera (6), different preset angles are used based on the angle with the horizontal direction of the transparent soil. Establish a multi-view triangulation model to measure planar displacement. Transformation into three-dimensional displacement in world coordinate system The formula is ； ； L is the baseline distance between the two cameras, and f is the focal length of the industrial camera. Planar displacement from the perspective of view Planar displacement from the perspective of view Planar displacement from the perspective of view This is the refractive index correction factor.

10. The image processing method for a transparent soil visualization test device according to claim 9, characterized in that, Based on the three-dimensional displacement field, the soil strain field and the quantitative index of borehole disturbance are further derived. The specific formulas and steps are as follows: P1. Calculate the spatial gradient of the smoothed three-dimensional displacement field to obtain the normal strain. With shear strain , , ,in These correspond to the horizontal radial deformation, the radial deformation of the borehole, the horizontal circumferential deformation, the circumferential deformation of the borehole, and the axial deformation of the borehole in the vertical depth direction, respectively. This represents the change in displacement for every 1 mm change in spatial position along the x-direction. These represent the shear strain in the xy plane and the shear strain in the xz plane, respectively. P2, Radius affected by borehole disturbance , The spatial coordinates of the drill bit center. The coordinates of feature points whose displacement exceeds a threshold value of 0.1 mm. In pixel equivalents; P3. The controller automatically generates a three-dimensional displacement field cloud map, strain distribution curve, and disturbance influence radius data, and associates them with the drill bit drilling angle and mud three parameters to form a quantitative test report.

Images