Slope construction quality detection method and system based on machine vision

Abstract

The invention relates to the technical field of side slope monitoring, in particular to a side slope construction quality detection method and system based on machine vision, and the method comprises the steps: continuously collecting speckle images of a side slope monitoring region, forming a region image sequence, and obtaining a displacement field and a strain field of the monitoring region through analyzing the local region image sequence; continuously collecting images of a plurality of key points in the monitoring area to form a key point image sequence, analyzing the key point image sequence to obtain displacement vectors of the key points, and correcting a displacement field of the local area through the displacement vectors of the key points; carrying out slope displacement early warning based on the strain field of the monitoring area and the corrected displacement field; and judging the slope quality based on the risk probability and the strain field of the local area.

Description

Technical Field

[0001] This invention relates to the field of slope monitoring technology, and in particular to a method and system for detecting slope construction quality based on machine vision. Background Technology

[0002] Monitoring the construction quality of road slopes is a crucial step in ensuring the safety, stability, and economy of slope engineering. The main monitoring content includes slope surface displacement, structural deformation, and cracks. Applying machine vision to slope displacement monitoring represents the forefront of current slope displacement monitoring technology. Essentially, it deeply integrates traditional surveying and computer vision to achieve non-contact, large-scale, high-frequency, and low-cost acquisition of deformation information.

[0003] In actual monitoring, how to combine point monitoring with area monitoring to overcome the sensitivity of traditional area monitoring (DIC monitoring based on digital image correlation) to environmental lighting conditions and the limitations of point monitoring range are problems that machine vision urgently needs to solve in slope displacement monitoring applications. Summary of the Invention

[0004] This invention performs point and surface monitoring on slopes simultaneously to obtain corresponding displacement data, and uses the point monitoring results (displacement vectors of key points) to correct the surface monitoring results (displacement field of the monitoring area), thereby improving the robustness and accuracy of slope displacement monitoring.

[0005] The technical solution proposed in this invention is: a method for inspecting the construction quality of slopes based on machine vision, the method comprising:

[0006] Speckle images of the slope monitoring area are continuously acquired to form a regional image sequence. By analyzing the local regional image sequence, the displacement field and strain field of the monitoring area are obtained.

[0007] Images of multiple key points within the monitoring area are continuously acquired to form a key point image sequence. By analyzing the key point image sequence, the displacement vectors of the key points are obtained, and the displacement field of the local area is corrected by the displacement vectors of the key points.

[0008] Slope displacement early warning is based on the strain field and corrected displacement field of the monitoring area;

[0009] The quality of the slope is determined based on the risk probability and the strain field in the local area.

[0010] Preferably, the continuous acquisition of speckle images of the slope monitoring area constitutes a regional image sequence. By analyzing the local regional image sequence, the displacement field and strain field of the monitoring area are obtained, including:

[0011] During the monitoring period, speckle images of the monitoring area are acquired at a preset acquisition frequency to form a regional image sequence. ;in, express Speckle images acquired at any time express The speckle images are collected at any time, and the monitoring period is... ;

[0012] by As a reference image, the monitoring area is divided into multiple grids on the reference image;

[0013] Each node of the grid is used as the center, i.e., the calculation point, and a square region is defined as a sub-region.

[0014] Assuming the slope deformation is uniform, let the speckle image acquired at the current moment be... , subregion The coordinates of the center are ;

[0015] The mapping equation is then: ;in, , Subregion The center pixel shift amount; , , , Subregion The first-order displacement gradient;

[0016] By solving the mapping equation using the Fast Fourier Transform, we obtain... ;

[0017] Then, an iterative optimization algorithm is used to solve the mapping equation to obtain the similarity. Maximum deformation parameter vector ;

[0018] in, ; The center of the reference image is indicated by The grayscale value of the sub-region; Indicates the center is The grayscale value of the sub-region; This represents the average grayscale value of all sub-regions in the reference image; express The average gray value of all sub-regions;

[0019] Will Transform the pixel coordinates of any node to the corresponding 3D space coordinates, including:

[0020] set up ,in, Indicates the scale factor. This represents the intrinsic parameter matrix of the camera used to acquire speckle images. This represents the extrinsic parameter matrix of the camera used to acquire speckle images; Represents three-dimensional spatial coordinates;

[0021] Solving the above equation using the least squares method yields the following results. Then, the three-dimensional displacement vector of any node is ;

[0022] in, Represents the spatial three-dimensional coordinates of the corresponding node on the reference image;

[0023] The displacement vector field of the monitoring area is The strain field is ;in, Indicates the number of sub-regions.

[0024] Preferably, the continuous acquisition of images of multiple key points within the monitoring area constitutes a key point image sequence, including:

[0025] Multiple key points are set up within the monitoring area; images of these key points are acquired at a preset acquisition frequency. ,in, ;in, Indicates the number of key points;

[0026] A sequence of key point images was obtained within the monitoring time period. .

[0027] Preferably, obtaining the displacement vectors of key points by analyzing the key point image sequence includes:

[0028] After converting the keypoint image to grayscale, the Hough circle transform is used to monitor the outer circle, obtaining... ID;

[0029] extract edge point set ,in, ; Indicates the number of edge points;

[0030] Fitting Ellipse Equation The center coordinates are obtained through geometric calculations. ;

[0031] time The center coordinates of all keypoint images constitute the time. The central coordinate set;

[0032] Transform the center coordinates to the corresponding three-dimensional space coordinates, including:

[0033] ,in, 'Indicates the scale factor, ' represents the intrinsic parameter matrix of the camera used to acquire keypoint images. This represents the extrinsic parameter matrix of the camera used to acquire keypoint images; Three-dimensional spatial coordinates representing the center coordinates;

[0034] The above equations are obtained by solving the least squares method. ;

[0035] Obtain the center coordinates at the initial time. Substitute The solution is obtained by the least squares method. Then the displacement vector of the key point is .

[0036] Preferably, the correction of the displacement field of a local region using the displacement vector of key points includes:

[0037] Register the 3D displacement vector of any node with the displacement vector of the key point in the coordinate system, including:

[0038] get and Transform by rotation, translation and scaling and Transformation to the same coordinate system includes:

[0039] Let the rotation transformation matrix be The translation transformation matrix is Scaling transformation ;

[0040] Constructing the objective equation ; where coordinates ;coordinate ;

[0041] By using the Protodyakonov analysis algorithm to solve the objective equation, we obtain... , , ;

[0042] Therefore, for any point within the monitored area image... Three-dimensional spatial coordinates All can be converted to In the coordinate system in which it is located, that is The transformed three-dimensional spatial coordinates are ;

[0043] Then, for the key points Calculate displacement residuals ; Key points within the monitored area image The transformed three-dimensional spatial coordinates of the location;

[0044] The corrected displacement field of the monitoring area is .

[0045] Preferably, the slope displacement early warning based on the strain field and corrected displacement field of the monitoring area includes:

[0046] Get and ,from and Several key indicators were extracted, including:

[0047] Resultant displacement Resultant velocity ; combined acceleration ,; strain localization factor ;

[0048] in, This represents the maximum theoretical shear strain (,) of a sub-region within the monitoring area. This represents the average value of the maximum theoretical shear strain in all sub-regions within the monitoring period; Represents the shear strain gradient The average value of the modulus;

[0049] For the center of each sub-area within the monitored area, determine whether the following conditions are met: and and ;in, , , This represents the resultant velocity threshold, resultant acceleration threshold, and strain localization factor threshold.

[0050] For all sub-regions that meet the above conditions, cluster analysis is performed using the DBSCAN algorithm within a time window to obtain multiple risk entities. ;

[0051] Extract evidence vectors for each risk subject ;in, This represents the average combined velocity of all subregions within the risk body. This represents the average combined acceleration across all subregions within the risk domain. This represents the average localization factor across all subregions within the risk body;

[0052] In a given state Downward observation of evidence vector The probability is ;in, Indicates mixed weights, Indicates the first The mean vector of Gaussian components; Represent the covariance matrix; ;in Indicates a stable state. This indicates the initial creep state. Indicates accelerated creep state. Indicates a state of destruction; Indicates the number of Gaussian components;

[0053] Let the overall risk probability be set. ;in, Indicates the reduction factor;

[0054] Furthermore, considering the future period of time Comprehensive risk probability within ;in, ,in, Representing state The risk factor;

[0055] Set risk probability thresholds and multiple risk probability ranges, based on Relationship with risk probability threshold and Based on the corresponding risk probability range, Level 1, Level 2, Level 3, and Level 4 warnings are generated, along with corresponding action recommendations.

[0056] Preferably, the slope displacement early warning based on the strain field and corrected displacement field of the monitoring area further includes:

[0057] Calculate the confidence level of the early warning. Among them, distribution entropy ; ;

[0058] if If so, the above warning levels and corresponding action recommendations shall be adopted;

[0059] If 0.5 If so, then the warning level will be lowered by one level and adopted;

[0060] if In this case, only the Level 1 warning level and corresponding action recommendations will be adopted.

[0061] Preferably, the determination of slope quality based on risk probability and local strain field includes:

[0062] Get and ;

[0063] Calculate the strain uniformity index ;in, This represents the standard deviation of strain across all sub-zones within the monitoring area; This represents the average strain across all sub-regions within the monitoring area; This represents a constant term used to avoid the numerator being zero; it is used to calculate the crack development index. Among them, the area of ​​the largest principal strain field ; This represents the critical strain at which the material cracks. Indicates an indicator function, when The value is 1 if the condition is met, and 0 otherwise. This represents the square of the strain gradient, used to distinguish between uniform high strain and cracks; Indicates the area of ​​the subregion; This represents the gradient sensitivity parameter; ;

[0064] Calculate the risk probability growth rate ;

[0065] The integrity score is calculated using the strain uniformity index, crack development index, and risk probability growth rate. ;in, , , Indicates the integrity weighting coefficient;

[0066] Dynamically update the integrity weight coefficient: ; ; ;in, , , This represents the integrity weight coefficient for the next time step from the current time step. Indicates a time interval;

[0067] if If so, the slope construction quality is judged to be excellent;

[0068] if If so, the slope construction quality is judged to be good;

[0069] if If so, the slope construction quality is judged to be moderate;

[0070] if If so, the slope construction quality is deemed dangerous.

[0071] A machine vision-based slope construction quality inspection system is provided, the system being used to execute the aforementioned machine vision-based slope construction quality inspection method.

[0072] A computer-readable storage medium storing a computer program that is executed by a processor to implement the aforementioned machine vision-based slope construction quality inspection method.

[0073] The beneficial effects of this invention are:

[0074] 1. This invention combines key displacement monitoring with regional displacement field monitoring, utilizing displacement residuals to correct the displacement field of the monitored area, thereby improving the accuracy of the displacement field. It not only solves the problem of difficulty in integrating monitoring accuracy and monitoring range in traditional technologies, but also achieves overall perception of slope surface displacement and focus on the displacement of key points.

[0075] 2. This invention provides slope displacement early warning based on the strain field and corrected displacement field of the monitoring area. It uses multi-dimensional evidence such as resultant displacement, resultant velocity, resultant acceleration and localization factor to perform comprehensive risk probability calculation and graded early warning, rather than based on single-point threshold comparison, making the early warning more reliable. Attached Figure Description

[0076] Figure 1 This is a flowchart of a slope construction quality inspection method based on machine vision according to the present invention. Detailed Implementation

[0077] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious modifications will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0078] It is understood that the term "a" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple, and the term "a" should not be understood as a limitation on the number.

[0079] refer to Figure 1 The technical solution provided by this invention is: a method for inspecting the construction quality of slopes based on machine vision, the method comprising:

[0080] Step 1: Continuously acquire speckle images of the slope monitoring area to form a regional image sequence. By analyzing the local regional image sequence, obtain the displacement field and strain field of the monitoring area. This specifically includes the following steps:

[0081] During the monitoring period, speckle images of the monitoring area are acquired at a preset acquisition frequency to form a regional image sequence. ;in, express Speckle images acquired at any time express The speckle images are collected at any time, and the monitoring period is... ;

[0082] by As a reference image, the monitoring area is divided into multiple grids on the reference image;

[0083] Define a square region (e.g., 31×31 pixels) as a sub-region, with each node of the grid as the center, i.e., the calculation point;

[0084] Assuming the slope deformation is uniform, let the speckle image acquired at the current moment be... , subregion The coordinates of the center are ;

[0085] The mapping equation is then: ;in, , Subregion The center pixel shift amount; , , , Subregion The first-order displacement gradient;

[0086] By solving the mapping equation using the Fast Fourier Transform, we obtain... ;

[0087] Then, an iterative optimization algorithm is used to solve the mapping equation to obtain the similarity. Maximum deformation parameter vector ;

[0088] in, ; The center of the reference image is indicated by The grayscale value of the sub-region; Indicates the center is The grayscale value of the sub-region; This represents the average grayscale value of all sub-regions in the reference image; express The average gray value of all sub-regions;

[0089] Will Transform the pixel coordinates of any node to the corresponding 3D space coordinates, including:

[0090] set up ,in, Indicates the scale factor. This represents the intrinsic parameter matrix of the camera used to acquire speckle images. This represents the extrinsic parameter matrix of the camera used to acquire speckle images; Represents three-dimensional spatial coordinates;

[0091] Solving the above equation using the least squares method yields the following results. Then, the three-dimensional displacement vector of any node is ;

[0092] in, Represents the spatial three-dimensional coordinates of the corresponding node on the reference image;

[0093] The displacement vector field of the monitoring area is The strain field is ;in, Indicates the number of sub-regions.

[0094] In this embodiment, the monitoring area is selected as a dangerous or typical area of ​​the slope, such as the top of the slope or the exposed weak interlayer. A high-contrast speckle pattern is sprayed on the monitoring area. For example, a layer of white primer is first sprayed on the surface of the monitoring area, and then black dots are sprayed on the white primer. The speckle image is obtained by acquiring the monitoring area with the speckle pattern through a high-resolution camera.

[0095] Step 2: Continuously acquire images of multiple key points within the monitoring area to form a key point image sequence. Analyze the key point image sequence to obtain the displacement vectors of the key points, and use these displacement vectors to correct the displacement field of the local area. This specifically includes the following steps:

[0096] Step 2.1: Set up multiple key points within the monitoring area. In this embodiment, manually coded targets are deployed within the monitoring area, and the locations of the manually coded targets are the key points.

[0097] Acquire key point images according to the preset acquisition frequency. ,in, ;in, Indicates the number of key points;

[0098] A sequence of key point images was obtained within the monitoring time period. .

[0099] Step 2.2: After converting the key point image to grayscale, use Hough circle transform to monitor the outer circle and obtain... ID;

[0100] extract edge point set ,in, ; Indicates the number of edge points;

[0101] Fitting Ellipse Equation The center coordinates are obtained through geometric calculations. ;

[0102] time The center coordinates of all keypoint images constitute the time. The central coordinate set;

[0103] Transform the center coordinates to the corresponding three-dimensional space coordinates, including:

[0104] ,in, 'Indicates the scale factor, ' represents the intrinsic parameter matrix of the camera used to acquire keypoint images. This represents the extrinsic parameter matrix of the camera used to acquire keypoint images; Three-dimensional spatial coordinates representing the center coordinates;

[0105] The above equations are obtained by solving the least squares method. ;

[0106] Obtain the center coordinates at the initial time. Substitute The solution is obtained by the least squares method. Then the displacement vector of the key point is .

[0107] Step 2.3: Register the 3D displacement vector of any node with the displacement vector of the key point in the coordinate system, including:

[0108] get and Transform by rotation, translation and scaling and Transformation to the same coordinate system includes:

[0109] Let the rotation transformation matrix be The translation transformation matrix is Scaling transformation ;

[0110] Constructing the objective equation ; where coordinates ;coordinate ;

[0111] By using the Protodyakonov analysis algorithm to solve the objective equation, we obtain... , , In slope monitoring, it is generally believed that... .

[0112] For any point within the monitored area image Three-dimensional spatial coordinates All can be converted to In the coordinate system in which it is located, that is The transformed three-dimensional spatial coordinates are ;

[0113] Then, for the key points Calculate displacement residuals ; Key points within the monitored area image The transformed three-dimensional spatial coordinates of the location;

[0114] The corrected displacement field of the monitoring area is .

[0115] Step 3: Based on the strain field and corrected displacement field of the monitoring area, perform slope displacement early warning, specifically including:

[0116] Get and ,from and Several key indicators were extracted, including:

[0117] Resultant displacement Resultant velocity ; combined acceleration In this embodiment, the above calculations need to be performed for each calculation point. Strain localization factor. ;

[0118] in, This represents the maximum theoretical shear strain of a sub-region within the monitored area (historical database or laboratory calibration). This represents the average value of the maximum theoretical shear strain in all sub-regions within the monitoring period; Represents the shear strain gradient The average value of the modulus;

[0119] For the center of each sub-area within the monitored area, determine whether the following conditions are met: and and ;in, , , This represents the resultant velocity threshold, resultant acceleration threshold, and strain localization factor threshold.

[0120] For all sub-regions that meet the above conditions, cluster analysis is performed using the DBSCAN algorithm within a time window to obtain multiple risk entities. ;

[0121] Extract evidence vectors for each risk subject ;in, This represents the average combined velocity of all subregions within the risk body. This represents the average combined acceleration across all subregions within the risk domain. This represents the average localization factor across all subregions within the risk body;

[0122] In a given state Downward observation of evidence vector The probability is Here we assume that in state The evidence vector distribution can be represented by a mixture of M Gaussian components. Among them, Indicates mixed weights, Indicates the first The mean vector of Gaussian components; Represent the covariance matrix; ;in Indicates a stable state. This indicates the initial creep state. Indicates accelerated creep state. Indicates a state of destruction; Indicates the number of Gaussian components;

[0123] Let the overall risk probability be set. ;in, This represents the reduction factor, which is taken as 0.3-0.5 in this embodiment. This formula indicates that even if the slope is at... There is also a certain probability that it will quickly evolve into a destructive state.

[0124] Furthermore, considering the future period of time Comprehensive risk probability within ;in, ,in, Representing state The risk factor, in this embodiment , , , .

[0125] if and If this occurs, a Level 1 warning will be triggered, and the following output will be displayed. ;in, This indicates the risk probability threshold; the recommended action is to increase the monitoring frequency and conduct manual verification.

[0126] if and If this occurs, a level two warning will be triggered, and output will be provided. Action recommendations; the action recommendations are: conduct on-site inspections and prepare contingency plans.

[0127] if and This will trigger a Level 3 warning and output... Action recommendations; the action recommendation is: activate the emergency response plan.

[0128] if and This will trigger a Level 4 warning and output [the warning]. Action recommendations; the action recommendation is: to take emergency measures.

[0129] In some preferred embodiments, this step further includes:

[0130] Calculate the confidence level of the early warning. Among them, distribution entropy ; ;

[0131] if If so, the above warning levels and corresponding action recommendations shall be adopted;

[0132] If 0.5 If so, then the warning level will be lowered by one level and adopted;

[0133] if In this case, only the Level 1 warning level and corresponding action recommendations will be adopted.

[0134] Step 4: Based on the risk probability and the strain field of the local area, determine the slope quality, which includes the following steps:

[0135] Get and ; Calculate the strain uniformity index ;in, This represents the standard deviation of strain across all sub-zones within the monitoring area; This represents the average strain across all sub-regions within the monitoring area; This represents a constant term, with a value of 10. -6 This is used to avoid the numerator being zero. A value close to 1 indicates uniform strain distribution and structural integrity, while a value close to 0 indicates highly concentrated strain and the presence of defects.

[0136] Calculate the crack development index Among them, the area of ​​the largest principal strain field ; This indicates the critical strain at which the material cracks (determined based on the type of soil and rock). Indicates an indicator function, when The value is 1 if the condition is met, and 0 otherwise. This represents the square of the strain gradient, used to distinguish between uniform high strain and cracks; Indicates the area of ​​the subregion; This represents the gradient sensitivity parameter; ;

[0137] Calculate the risk probability growth rate ;

[0138] The integrity score is calculated using the strain uniformity index, crack development index, and risk probability growth rate. ;in, , , Indicates the integrity weighting coefficient;

[0139] Dynamically update the integrity weight coefficient:

[0140] ;

[0141] ;

[0142] ;in, , , This represents the integrity weight coefficient for the next time step from the current time step. Indicates a time interval;

[0143] if If the slope construction quality is good, then routine monitoring is sufficient.

[0144] if If the slope construction quality is good, then it is necessary to strengthen inspections.

[0145] if If the slope construction quality is deemed to be moderate, then special testing or reinforcement is required.

[0146] if If the slope construction quality is deemed dangerous, immediate emergency reinforcement is required.

[0147] The present invention also provides a machine vision-based slope construction quality inspection system, which is used to perform the aforementioned machine vision-based slope construction quality inspection method.

[0148] The present invention also provides a computer-readable storage medium storing a computer program that is executed by a processor to implement the aforementioned machine vision-based slope construction quality inspection method.

[0149] The processes described above with reference to the flowcharts in the embodiments disclosed in this invention can be implemented as computer software programs. The embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), it performs the functions defined in the methods of this application. It should be noted that the computer-readable medium described above in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wire segments, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless segments, wire segments, optical fibers, RF, etc., or any suitable combination thereof.

[0150] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation that may be implemented in systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0151] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved, and the functions and structural principles of the present invention have been demonstrated and explained in the embodiments. Without departing from the stated principles, the implementation of the present invention may have any changes or modifications.

Claims

1. A machine vision-based method for detecting construction quality of a slope, characterized in that, The method includes: Speckle images of the slope monitoring area are continuously acquired to form a regional image sequence. By analyzing the local regional image sequence, the displacement field and strain field of the monitoring area are obtained. Images of multiple key points within the monitoring area are continuously acquired to form a key point image sequence. By analyzing the key point image sequence, the displacement vectors of the key points are obtained, and the displacement field of the local area is corrected using the displacement vectors of the key points. Slope displacement early warning is based on the strain field and corrected displacement field of the monitoring area; The quality of the slope is determined based on the risk probability and the strain field in the local area.

2. The method for inspecting slope construction quality based on machine vision according to claim 1, characterized in that, The continuously acquired speckle images of the slope monitoring area constitute a regional image sequence. By analyzing the local regional image sequence, the displacement field and strain field of the monitoring area are obtained, including: During the monitoring period, speckle images of the monitoring area are acquired at a preset acquisition frequency to form a regional image sequence. ;in, express Speckle images acquired at any time express The speckle images are collected at any time, and the monitoring period is... ; by As a reference image, the monitoring area is divided into multiple grids on the reference image; Each node of the grid is used as the center, i.e., the calculation point, and a square region is defined as a sub-region. Assuming the slope deformation is uniform, let the speckle image acquired at the current moment be... , subregion The coordinates of the center are ; The mapping equation is then: ;in, , Subregion The center pixel shift amount; , , , Subregion The first-order displacement gradient; By solving the mapping equation using the Fast Fourier Transform, we obtain... ; Then, an iterative optimization algorithm is used to solve the mapping equation to obtain the similarity. Maximum deformation parameter vector ; in, The center of the reference image is indicated by The grayscale value of the sub-region; Indicates the center is The grayscale value of the sub-region; This represents the average grayscale value of all sub-regions in the reference image; express The average grayscale value of all sub-regions; Will Transform the pixel coordinates of any node to the corresponding 3D space coordinates, including: set up ,in, Indicates the scale factor. This represents the intrinsic parameter matrix of the camera used to acquire speckle images. This represents the extrinsic parameter matrix of the camera used to acquire speckle images; Represents three-dimensional spatial coordinates; Solving the above equation using the least squares method yields the following results. Then, the three-dimensional displacement vector of any node is ； in, Represents the spatial three-dimensional coordinates of the corresponding node on the reference image; The displacement vector field of the monitoring area is The strain field is ;in, Indicates the number of sub-regions.

3. The method for inspecting slope construction quality based on machine vision according to claim 2, characterized in that, The continuous acquisition of images of multiple key points within the monitoring area constitutes a key point image sequence, including: Multiple key points are set up within the monitoring area; images of these key points are acquired at a preset acquisition frequency. ,in, ;in, Indicates the number of key points; A sequence of key point images was obtained within the monitoring time period. .

4. The method for inspecting slope construction quality based on machine vision according to claim 3, characterized in that, The step of obtaining the displacement vectors of key points by analyzing the key point image sequence includes: After converting the keypoint image to grayscale, the Hough circle transform is used to monitor the outer circle, obtaining... ID; extract edge point set ,in, ; Indicates the number of edge points; Fitting Ellipse Equation ， The center coordinates were obtained through geometric calculations. ; time The center coordinates of all keypoint images constitute the time. The central coordinate set; Transform the center coordinates to the corresponding three-dimensional space coordinates, including: ,in, 'Indicates the scale factor, ' represents the intrinsic parameter matrix of the camera used to acquire keypoint images. This represents the extrinsic parameter matrix of the camera used to acquire keypoint images; Three-dimensional spatial coordinates representing the center coordinates; The above equations are obtained by solving the least squares method. ; Obtain the center coordinates at the initial time. Substitute ， The solution is obtained by solving the least squares method. Then the displacement vector of the key point is 。 5. The method for inspecting slope construction quality based on machine vision according to claim 4, characterized in that, The method of correcting the displacement field of a local region using the displacement vector of key points includes: Register the 3D displacement vector of any node with the displacement vector of the key point in the coordinate system, including: get and Transform by rotation, translation and scaling and Transformation to the same coordinate system includes: Let the rotation transformation matrix be The translation transformation matrix is Scaling transformation ; Constructing the objective equation ； Where, coordinates ;coordinate ; By using the Protodyakonov analysis algorithm to solve the objective equation, we obtain... , , ; Therefore, for any point within the monitored area image... Three-dimensional spatial coordinates All can be converted to In the coordinate system in which it is located, that is The transformed three-dimensional spatial coordinates are ； Then, for the key points Calculate displacement residuals ; Key points within the monitored area image The transformed three-dimensional spatial coordinates of the location; The corrected displacement field of the monitoring area is 。 6. The method for inspecting slope construction quality based on machine vision according to claim 5, characterized in that, The slope displacement early warning based on the strain field and corrected displacement field of the monitoring area includes: Get and ,from and Several key indicators were extracted, including: Resultant displacement Resultant velocity ; combined acceleration ,; strain localization factor ; in, This represents the maximum theoretical shear strain in a sub-region within the monitored area. This represents the average value of the maximum theoretical shear strain in all sub-regions within the monitoring period; Represents the shear strain gradient The average value of the modulus; For the center of each sub-area within the monitored area, determine whether the following conditions are met: and and ;in, , , This represents the resultant velocity threshold, resultant acceleration threshold, and strain localization factor threshold. For all sub-regions that meet the above conditions, cluster analysis is performed using the DBSCAN algorithm within a time window to obtain multiple risk entities. ; Extract evidence vectors for each risk subject ;in, This represents the average combined velocity of all subregions within the risk body. This represents the average combined acceleration across all subregions within the risk domain. This represents the average localization factor across all subregions within the risk body; In a given state Downward observation of evidence vector The probability is ;in, Indicates mixed weights, Indicates the first The mean vector of Gaussian components; Represent the covariance matrix; ;in Indicates a stable state. This indicates the initial creep state. Indicates accelerated creep state. Indicates a state of destruction; Indicates the number of Gaussian components; Let the overall risk probability be set. ;in, Indicates the reduction factor; Furthermore, considering the future period of time Comprehensive risk probability within ;in, ,in, Representing state The risk factor; Set risk probability thresholds and multiple risk probability ranges, based on Relationship with risk probability threshold and Based on the corresponding risk probability range, Level 1, Level 2, Level 3, and Level 4 warnings are generated, along with corresponding action recommendations.

7. The method for inspecting slope construction quality based on machine vision according to claim 6, characterized in that, The slope displacement early warning based on the strain field and corrected displacement field of the monitoring area also includes: Calculate the confidence level of the early warning. Among them, distribution entropy ; ; if If so, the above warning levels and corresponding action recommendations shall be adopted; If 0.5 If so, then the warning level will be lowered by one level and adopted; if In this case, only the Level 1 warning level and corresponding action recommendations will be adopted.

8. The method for inspecting slope construction quality based on machine vision according to claim 7, characterized in that, The method of determining slope quality based on risk probability and local strain field includes: Get and ； Calculate the strain uniformity index ;in, This represents the standard deviation of strain across all sub-zones within the monitoring area; This represents the average strain across all sub-regions within the monitoring area; This represents a constant term used to avoid the numerator being zero; it is used to calculate the crack development index. Among them, the area of ​​the largest principal strain field ; This represents the critical strain at which the material cracks. Indicates an indicator function, when The value is 1 if the condition is met, and 0 otherwise. This represents the square of the strain gradient, used to distinguish between uniform high strain and cracks; Indicates the area of ​​the subregion; This represents the gradient sensitivity parameter; ; Calculate the risk probability growth rate ; The integrity score is calculated using the strain uniformity index, crack development index, and risk probability growth rate. ;in, , , Indicates the integrity weighting coefficient; Dynamically update the integrity weight coefficient: ； ； ； in, , , This represents the integrity weight coefficient for the next time step from the current time step. Indicates a time interval; if If so, the slope construction quality is judged to be excellent; if If so, the slope construction quality is judged to be good; if If so, the slope construction quality is judged to be moderate; if If so, the slope construction quality is deemed dangerous.

9. A slope construction quality inspection system based on machine vision, characterized in that, The system is used to perform a machine vision-based slope construction quality inspection method as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is executed by a processor to implement a machine vision-based slope construction quality inspection method as described in any one of claims 1-8.

Images