Tunnel rock mass expansion monitoring and measurement method and system

Abstract

The present invention relates to the field of tunnel technology, specifically disclosing a method and system for monitoring and measuring tunnel rock expansion. The system includes a bracket, a first electric pan-tilt head mounted on the bracket, and a shooting module mounted on the navigation system of the first electric pan-tilt head. The system also includes a control module, an acquisition module, and an edge processing module. The control module is configured to control the rotation of the first electric pan-tilt head according to a preset rotation trajectory. The acquisition module is configured to control the shooting module to capture a local image of the tunnel face after the first electric pan-tilt head reaches a predetermined position. The acquisition module is configured to obtain the local image from the shooting module and send the local image to the edge processing module. The edge processing module is configured to analyze the local image according to a deployed AI model and output a rock mass state recognition result based on the analysis results of all local images of the tunnel face. The technical solution of the present invention can automatically complete tunnel face photography and analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnels, and in particular to a method and system for monitoring and measuring tunnel body rock mass expansion. Background Art

[0002] Tertiary expansive mudstone contains a high concentration of hydrophilic minerals such as montmorillonite and illite, which soften and expand upon contact with water, shrinking and cracking as water is lost. Tunnels constructed in expansive rock often experience rapid, destructive, prolonged deformation, and difficulty in remediation. During construction, initial support deformation often increases, leading to poor water stability. This, in turn, can lead to lining cracking, inverted arch floor heave, and roadbed uplift.

[0003] Therefore, during construction, it is necessary to meet basic monitoring and measurement requirements and conduct targeted monitoring and measurement of the surrounding rock to determine its stability, understand the deformation characteristics of the surrounding rock and support structure, and determine the timing of anchoring, support, and secondary lining application, so as to timely adjust support parameters and construction methods. Only by excavating the geological conditions exposed can we objectively understand the surrounding rock problems. Therefore, evaluating the dynamics of the surrounding rock based on the geological conditions exposed at the face becomes the key to construction.

[0004] Monitoring and measurement can be divided into mandatory and optional items. Mandatory items include evaluating the condition of the tunnel face surrounding rock, observing deformation and cracking on the support surface, measuring horizontal displacement, measuring vault subsidence, and measuring displacement of the arch foot. For evaluating the condition of the tunnel face surrounding rock, a tripod must be set up at the deformation measurement point. A digital camera is placed on the tripod to photograph and record the tunnel face, and the corresponding working face observation form and construction phase surrounding rock grade determination card are completed. The entire photography and analysis process is completely manual, and these operations need to be repeated during tunnel excavation, which is very labor-intensive.

[0005] Therefore, a method and system for monitoring and measuring tunnel rock expansion that can automatically complete tunnel face photography and analysis is needed. Summary of the Invention

[0006] One of the purposes of the present invention is to provide a tunnel rock expansion monitoring and measurement system that can automatically complete tunnel face photography and analysis.

[0007] In order to solve the above technical problems, this application provides the following technical solutions:

[0008] The tunnel rock expansion monitoring and measurement system includes a bracket, a first electric pan-tilt head installed on the bracket, and a shooting module installed on the navigation of the first electric pan-tilt head; it also includes a control module, an acquisition module and an edge processing module; the control module is used to control the rotation of the first electric pan-tilt head according to a preset rotation trajectory; it is also used to control the shooting module to capture a local image of the tunnel face after the first electric pan-tilt head reaches a predetermined position; the acquisition module is used to obtain the local image from the shooting module and send the local image to the edge processing module; the edge processing module is used to analyze the local image according to the deployed AI model, and output the rock state recognition result based on the analysis results of all local images of the tunnel face.

[0009] The basic scheme principles and beneficial effects are as follows:

[0010] In this solution, the first motorized pan / tilt automatically adjusts the camera module's shooting angle to capture a partial image of the tunnel face. The AI ​​model deployed in the edge processing module then analyzes this partial image and automatically outputs rock mass state identification results, automating tunnel face capture and analysis. Using edge computing for image analysis is less affected by network conditions than cloud-based analysis, resulting in more timely analysis.

[0011] Furthermore, it also includes a communication module and a server; the communication module is also used to upload the rock state identification result to the server; the server is also used to generate a tunnel face observation record table according to the rock state identification result.

[0012] It is convenient for relevant personnel to access the server to view the results.

[0013] Furthermore, the invention also includes a lighting module and a second electric pan-tilt head, wherein the second electric pan-tilt head is fixed on the bracket, and the lighting module is fixed on the second electric pan-tilt head;

[0014] The edge processing module is also used to identify whether there is an underexposed area in the currently acquired local image. If there is an underexposed area, the underexposed information is sent to the control module;

[0015] The control module is also used to control the second electric pan-tilt head to rotate in accordance with the first electric pan-tilt head; and is also used to control the lighting module to turn on after receiving underexposure information.

[0016] Because tunnel construction often relies on artificial light sources with limited illumination, coupled with potential obstruction by construction equipment, insufficient brightness can occur in certain areas of the tunnel face, impacting image quality and subsequent analysis. This preferred solution uses a lighting module to fill in underexposed areas, ensuring image quality.

[0017] Furthermore, the edge processing module is further configured to send local image qualification information to the control module when there is no underexposed area in the local image;

[0018] The control module is further configured to control the first electric pan-tilt platform to rotate to a next predetermined position along a rotation trajectory after receiving the qualified information of the partial image.

[0019] Furthermore, the server is also used to receive the trained AI model and deploy the trained AI model to the edge processing module through the communication module.

[0020] Furthermore, the rock mass state identification results include water gushing conditions, weathering degree, geological structure influence degree and crack morphology.

[0021] A second object of the present invention is to provide a method for monitoring and measuring the expansion of a tunnel rock mass, comprising the following contents:

[0022] S1. Train the AI ​​model. The server receives the trained AI model and deploys it to the edge processing module through the communication module.

[0023] S2. Place a bracket at the measurement point and fix the shooting module and the lighting module on the first electric pan-tilt platform and the second electric pan-tilt platform respectively;

[0024] S3, controlling the first electric pan-tilt platform to rotate according to a preset rotation trajectory; after the first electric pan-tilt platform reaches a predetermined position, controlling the shooting module to capture a partial image of the palm face; and controlling the second electric pan-tilt platform to rotate in accordance with the first electric pan-tilt platform;

[0025] S4, the acquisition module obtains the local image from the shooting module and sends the local image to the edge processing module;

[0026] S5. The edge processing module identifies whether there is an underexposed area in the currently acquired partial image. If so, the edge processing module sends underexposed information to the control module. After receiving the underexposed information, the control module controls the lighting module to turn on and controls the camera module to retake the partial image of the face.

[0027] If there is no underexposed area, the control module sends the qualified information of the partial image to the control module; after receiving the qualified information of the partial image, the control module controls the first electric pan / tilt head to rotate to the next predetermined position along the rotation trajectory;

[0028] S6. The edge processing module analyzes the local images according to the deployed AI model and outputs the rock mass state recognition results based on the analysis results of all local images of the tunnel face.

[0029] Furthermore, it also includes:

[0030] S7. The communication module uploads the rock mass state identification result to the server; the server is also used to generate a tunnel face observation record table based on the rock mass state identification result.

[0031] Furthermore, in step S6, the rock mass state identification result includes water gushing condition, weathering degree, geological structure influence degree and fracture morphology. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a logic block diagram of the first embodiment of the tunnel rock expansion monitoring and measurement system. DETAILED DESCRIPTION

[0033] The following is further described in detail through specific implementation methods:

[0034] Example 1

[0035] like Figure 1 As shown, the cave rock expansion monitoring and measurement system of this embodiment includes a bracket, a first electric pan-tilt head installed on the bracket, and a shooting module installed on the first electric pan-tilt head navigation; it also includes a control module, an acquisition module, an edge processing module, a communication module, a server and a lighting module.

[0036] The control module is configured to control the rotation of the first electric pan-tilt platform according to a preset rotation trajectory and to control the camera module to capture a partial image of the palm face after the first electric pan-tilt platform reaches a predetermined position. In this embodiment, the bracket is positioned at the measurement point, and the camera module utilizes a digital camera to pre-divide the palm face into four capture sections: lower left, lower right, upper right, and upper left. The digital camera is initially aligned with the palm face. The control module then controls the rotation of the first electric pan-tilt platform, directing the digital camera toward the lower left section to capture a partial image of the lower left section. The digital camera then directs the digital camera toward the upper left section to capture a partial image of the upper left section. The digital camera then directs the digital camera toward the upper right section to capture a partial image of the upper right section. Finally, the mobile digital camera directs the mobile digital camera toward the lower right section to capture a partial image of the lower right section. Adjacent partial images partially overlap. In other embodiments, the divisions of the palm face can also be determined based on the distance between the measurement point and the palm face and the focal length used by the digital camera during capture.

[0037] The acquisition module is used to obtain the local image from the shooting module and send the local image to the edge processing module;

[0038] The edge processing module is used to analyze local images according to the deployed AI model and output rock mass status recognition results based on the analysis results of all local images of the tunnel face.

[0039] The communication module is also used to upload the rock mass state identification results to the server; the server is also used to generate a tunnel face observation record table based on the rock mass state identification results. The rock mass state identification results include water inflow, weathering degree, geological structure influence, and fracture morphology. Water inflow includes none, seepage, overall wetness, gushing, or erupting; weathering degree includes unweathered, slightly weathered, weakly weathered, and strongly weathered; geological structure influence includes slight, severe, severe, and very severe; and fracture morphology includes no fractures, random, and directional. In this embodiment, the rock mass state identification results are entered into a preset blank tunnel face observation record table.

[0040] The server is also used to receive the trained AI model and deploy the trained AI model to the edge processing module through the communication module.

[0041] In this embodiment, model training requires collecting a large number of local images of different tunnel face rock mass conditions to ensure that the local images adequately cover various tunnel face rock mass conditions. These local images also require preprocessing, including image resizing and data annotation. The annotated images are then divided into training and test sets.

[0042] In this example, a convolutional neural network was used to construct the AI ​​model. During training, images from the training set were fed into the AI ​​model, and backpropagation was performed to optimize the model parameters based on the annotations. Finally, the performance of the trained model was evaluated using the test set. After the AI ​​model training was complete, it was uploaded to the server.

[0043] The second electric pan-tilt head is fixed on the bracket, and the lighting module is fixed on the second electric pan-tilt head. In this embodiment, the lighting module adopts an LED light source. The control module is also used to control the second electric pan-tilt head to rotate in accordance with the first electric pan-tilt head.

[0044] The edge processing module is also used to identify whether there is an underexposed area in the currently acquired local image. If there is an underexposed area, the underexposed area information is sent to the control module; the control module is also used to control the lighting module to turn on after receiving the underexposed area information.

[0045] The edge processing module is further configured to send local image qualification information to the control module when there is no underexposed area in the local image;

[0046] The control module is further configured to control the first electric pan / tilt platform to rotate to the next predetermined position along the rotation trajectory after receiving the qualified partial image information. In this embodiment, another trained AI model can be used to detect underexposed areas.

[0047] This embodiment also provides a method for monitoring and measuring the expansion of a tunnel rock mass, using the above-mentioned system, including the following contents:

[0048] S1. Train the AI ​​model. The server receives the trained AI model and deploys it to the edge processing module through the communication module.

[0049] S2. Place a bracket at the measurement point and fix the shooting module and the lighting module on the first electric pan-tilt platform and the second electric pan-tilt platform respectively;

[0050] S3, controlling the first electric pan-tilt platform to rotate according to a preset rotation trajectory; after the first electric pan-tilt platform reaches a predetermined position, controlling the shooting module to capture a partial image of the palm face; and controlling the second electric pan-tilt platform to rotate in accordance with the first electric pan-tilt platform;

[0051] S4, the acquisition module obtains the local image from the shooting module and sends the local image to the edge processing module;

[0052] S5. The edge processing module identifies whether there is an underexposed area in the currently acquired partial image. If so, the edge processing module sends underexposed information to the control module. After receiving the underexposed information, the control module controls the lighting module to turn on and controls the camera module to retake the partial image of the face.

[0053] If there is no underexposed area, the partial image qualified information is sent to the control module; after receiving the partial image qualified information, the control module controls the first electric pan-tilt head to rotate to the next predetermined position along the rotation trajectory.

[0054] S6: The edge processing module analyzes local images using the deployed AI model and outputs rock mass condition identification results based on the analysis results of all local images of the tunnel face. Rock mass condition identification results include water inflow, weathering degree, geological structure influence, and fracture morphology. Water inflow conditions include none, seepage, overall wetness, gushing, or erupting; weathering degree includes unweathered, slightly weathered, weakly weathered, and strongly weathered; geological structure influence includes slight, severe, severe, and very severe; and fracture morphology includes none, random, and directional.

[0055] S7. The communication module uploads the rock mass state identification result to the server; the server is also used to generate a tunnel face observation record table based on the rock mass state identification result.

[0056] Example 2

[0057] The difference between this embodiment and the first embodiment is that in this embodiment, the control module is further configured to control the shooting module to shoot a complete image of the tunnel face in the initial state; and the acquisition module is further configured to send the complete image to the edge processing module.

[0058] The edge processing module is also used to identify whether there are underexposed areas in the currently acquired complete image. If not, the edge processing module is also used to analyze the complete image based on the deployed AI model and output the rock state identification results.

[0059] If there are underexposed areas, the number of underexposed areas is calculated and it is determined whether the number is equal to a threshold; in this embodiment, the threshold is 1.

[0060] If it is not greater than the threshold, the edge processing module is used to determine the shooting part to which the underexposed area belongs, as well as the coordinates of the center point of the underexposed area in the image, and calculate the position coordinates of the corresponding center point on the palm face.

[0061] Specifically,

[0062] First, perform camera calibration: obtain the camera's intrinsic parameters (camera matrix) and extrinsic parameters (rotation matrix and translation vector). Camera calibration can be performed using a calibration plate or other calibration methods. Intrinsic parameters include focal length, principal point, image size, pixel size, and distortion parameters. Extrinsic parameters include rotation matrix and translation vector.

[0063] Then perform inverse projection: using the coordinates of the center point in the image (x_A, y_A) and the intrinsic parameters of the camera, perform an inverse projection operation to map the center point to a ray in the camera coordinate system.

[0064] a. Convert the coordinates of the center point to the values ​​of the normalized plane coordinate system:

[0065] u_A=(x_A-c_x) / f_x

[0066] v_A=(y_A-c_y) / f_y

[0067] b. Construct the ray direction vector:

[0068] ray_direction=[u_A,v_A,1]

[0069] c. Normalize the ray direction vector:

[0070] ray_direction_normalized = ray_direction / ||ray_direction||, where ||ray_direction|| is the magnitude of the ray direction vector;

[0071] Finally, three-dimensional reconstruction is performed: using the rays obtained by back projection, the intersection with the tunnel face is found in three-dimensional space.

[0072] a. Convert the ray from the camera coordinate system to the world coordinate system:

[0073] Calculate the coordinates of the starting point of the ray:

[0074] ray_origin = [0,0,0], where [0,0,0] represents the camera's position coordinates;

[0075] Transform the ray origin in the camera coordinate system (ray_origin) by the rotation matrix (R) and the translation vector (T) to obtain the ray origin in the world coordinate system (ray_origin_world):

[0076] ray_origin_world=R*ray_origin+T

[0077] b. Convert the ray direction from the camera coordinate system to the world coordinate system:

[0078] ray_direction_world=R*ray_direction_normalized

[0079] c. Intersection with the palm face:

[0080] Based on the geometry of the tunnel face, the coordinates of the intersection of the ray and the tunnel face are calculated. This is the coordinates of the center point on the tunnel face (X_w, Y_w, Z_w, where X_w and Y_w represent the horizontal and vertical positions on the tunnel face, and Z_w represents the depth position on the tunnel face). The geometry of the tunnel face includes the plane equation or triangle vertices of the tunnel face. In this embodiment, the plane equation is used.

[0081] The edge processing module is also used to analyze the complete image based on the deployed AI model, output the rock state identification results, and calculate the evaluation value based on the rock state identification results. In this embodiment, each type of situation corresponds to a different score, and the sum of the scores is the evaluation value. For example, in the case of water gushing, no score is 1 point, seepage is 2 points, overall wetness is 3 points, and gushing or erupting is 4 points; in other words, the more serious the situation, the higher the score.

[0082] The edge processing module is also used to determine whether the evaluation value falls within the first interval. If so, it sends an instruction to the control module to end shooting. Falling within the first interval indicates that the rock mass condition is good and no local image analysis is required, which can effectively save computing power and allow the edge processing module to handle other matters in tunnel construction more efficiently.

[0083] If it belongs to the second interval, an instruction to continue shooting is sent to the control module; the edge processing module is also used to upload the rock state recognition results and the corresponding local images to the server through the communication module; belonging to the second interval indicates that the rock condition is good, and local images continue to be shot but are uploaded directly without analysis and processing, which can be used for archiving. Since there are already recognition results, there is no need to consider the delay problem, and the server can also perform analysis and processing, saving the computing power of the edge processing module.

[0084] If the image falls within the third interval, a command to continue recording is sent to the control module. The edge processing module is further configured to analyze the local image using the deployed AI model and output a rock mass state identification result based on the analysis results of all local images of the tunnel face. In this embodiment, the evaluation value ranges for the first, second, and third intervals increase in increments. The specific values ​​can be set based on actual conditions, for example, 3-5 points for the first interval, 6-7 points for the second interval, and 8 points or above for the third interval.

[0085] During continued shooting, if the number of underexposed areas equals a threshold, the control module is further configured to control the rotation of the second motorized pan / tilt head based on the position coordinates of the center point on the palm face, before the shooting module captures a partial image of the portion corresponding to the underexposed area, align the lighting module with the center point on the palm face, and activate the lighting module. In this embodiment, the lighting module includes a telephoto LED light source and a wide-angle LED light source of different focal lengths. The lighting module is aligned with the center point on the palm face, and the telephoto LED light source is activated.

[0086] If the number of underexposed areas exceeds a threshold, the control module controls the second motorized pan / tilt head to follow the first motorized pan / tilt head. The edge processing module is also responsible for identifying whether any underexposed areas exist in the currently acquired partial image. If so, it sends underexposure information to the control module. Upon receiving this information, the control module is further responsible for activating the wide-angle LED light source. Activating the wide-angle LED light source provides wider coverage, providing fill light to both underexposed and normal areas.

[0087] The above are only embodiments of the present invention. The invention is not limited to the fields involved in this implementation case. Common knowledge such as the known specific structures and characteristics in the scheme is not described in detail here. Ordinary technicians in the relevant field are aware of all common technical knowledge in the technical field to which the invention belongs before the application date or priority date, can obtain all existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the inspiration given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several variations and improvements can be made, which should also be regarded as the scope of protection of the present invention. These will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A cave rock expansion monitoring and measurement system, comprising a bracket, a first electric pan-tilt head mounted on the bracket, and a camera module mounted on the first electric pan-tilt head; characterized in that: It also includes a control module, an acquisition module, an edge processing module, a communication module and a server; The control module is used to control the shooting module to shoot a complete image of the tunnel face in the initial state; the acquisition module is also used to send the complete image to the edge processing module; The edge processing module is used to identify whether there are underexposed areas in the currently acquired complete image. If not, the edge processing module is also used to analyze the complete image based on the deployed AI model, output the rock mass state recognition results, and calculate the evaluation value based on the rock mass state recognition results; The edge processing module is further configured to determine whether the evaluation value falls within the first interval, and if so, to send an instruction to the control module to end the shooting; If it belongs to the second interval, sending a command to the control module to continue shooting; If it is in the third interval, sending a command to the control module to continue shooting; The control module is used to control the rotation of the first electric pan-tilt platform according to a preset rotation trajectory; and is also used to control the shooting module to shoot a partial image of the palm face after the first electric pan-tilt platform reaches a predetermined position; The acquisition module is used to obtain the local image from the shooting module and send the local image to the edge processing module; If it belongs to the second interval, the edge processing module is further used to upload the rock state recognition result and the corresponding local image to the server through the communication module; If it falls within the third interval, the edge processing module is further configured to analyze the local image using the deployed AI model, output a rock mass state recognition result based on the analysis results of all local images of the tunnel face, and upload the rock mass state recognition result to the server; The server is also used to generate a tunnel face observation record sheet based on the rock mass state identification results; If there are underexposed areas, calculate the number of underexposed areas and determine whether the number is equal to the threshold; If it is not greater than the threshold; The edge processing module is used to determine the shooting part to which the underexposed area belongs, as well as the coordinates of the center point of the underexposed area in the image, and calculate the position coordinates of the corresponding center point on the tunnel face: Including: calibrating the shooting module to obtain the internal and external parameters of the shooting module. The internal parameters include focal length, principal point, image size, pixel size and distortion parameters; the external parameters include rotation matrix and translation vector; Perform inverse projection, using the coordinates of the center point in the image and the intrinsic parameters of the shooting module, to map the center point to a ray in the shooting module coordinate system; It also includes a lighting module and a second electric pan-tilt head, the second electric pan-tilt head is fixed on the bracket, and the lighting module is fixed on the second electric pan-tilt head; The edge processing module is also used to identify whether there is an underexposed area in the currently acquired local image. If there is an underexposed area, the underexposed information is sent to the control module; The control module is also used to control the second electric pan-tilt head to rotate in accordance with the first electric pan-tilt head; and is also used to control the lighting module to turn on after receiving underexposure information.

2. The cave rock expansion monitoring and measurement system according to claim 1, characterized in that: The edge processing module is further configured to send local image qualification information to the control module when there is no underexposed area in the local image; The control module is further configured to control the first electric pan-tilt platform to rotate to a next predetermined position along a rotation trajectory after receiving the qualified information of the partial image.

3. The cave rock expansion monitoring and measurement system according to claim 2, characterized in that: The server is also used to receive the trained AI model and deploy the trained AI model to the edge processing module through the communication module.

4. The cave rock expansion monitoring and measurement system according to claim 3 is characterized in that: The rock mass state identification results include water gushing conditions, weathering degree, geological structure influence degree and crack morphology.

5. A method for monitoring and measuring the expansion of a tunnel rock mass, using the system according to any one of claims 1 to 4, characterized in that: Includes the following: S1. Train the AI ​​model. The server receives the trained AI model and deploys it to the edge processing module through the communication module. S2. Place a bracket at the measurement point and fix the shooting module and the lighting module on the first electric pan-tilt platform and the second electric pan-tilt platform respectively; S3, controlling the first electric pan-tilt platform to rotate according to a preset rotation trajectory; after the first electric pan-tilt platform reaches a predetermined position, controlling the shooting module to capture a partial image of the palm face; and controlling the second electric pan-tilt platform to rotate in accordance with the first electric pan-tilt platform; S4, the acquisition module obtains the local image from the shooting module and sends the local image to the edge processing module; S5. The edge processing module identifies whether there is an underexposed area in the currently acquired partial image. If so, the edge processing module sends underexposed information to the control module. After receiving the underexposed information, the control module controls the lighting module to turn on and controls the camera module to retake the partial image of the face. If there is no underexposed area, the control module sends the qualified information of the partial image to the control module; after receiving the qualified information of the partial image, the control module controls the first electric pan / tilt head to rotate to the next predetermined position along the rotation trajectory; S6. The edge processing module analyzes the local images according to the deployed AI model and outputs the rock mass state recognition results based on the analysis results of all local images of the tunnel face.

6. The method for monitoring and measuring cave rock expansion according to claim 5, characterized in that: Also includes: S7, the communication module uploads the rock mass state identification result to the server; The server is also used to generate a tunnel face observation record sheet based on the rock mass state identification results.

7. The method for monitoring and measuring cave rock expansion according to claim 6, characterized in that: In step S6, the rock mass state identification result includes water gushing condition, weathering degree, geological structure influence degree and fracture morphology.

Images