A machine vision displacement calibration method for foundation pit construction

Abstract

The present invention provides a machine vision displacement calibration method for foundation pit construction, which belongs to the field of displacement calibration technology and includes the following steps: step 1: obtaining the structure and position of the target foundation pit construction and obtaining a corresponding camera; step 2: obtaining data on the distortion parameters and internal focus parameters of the camera and calibrating the camera; step 3: capturing image data of the foundation pit construction using the calibrated camera and performing preprocessing; step 4: extracting significant feature points at the foundation pit edge, predicting the significant feature points based on a support vector machine, and determining the actual position of the foundation pit; step 5: comparing the actual position with the original set position, obtaining the displacement deviation, and performing displacement calibration to achieve accurate correction of the foundation pit position displacement. The method solves the problem of low automation level in the process of using a fixed reference point for point guidance, the need for direct contact with the detection object, and the instability of the instrument caused by long-term operation, which reduces the accuracy of the result of determining the foundation pit offset state.

Description

Technical Field

[0001] The present invention relates to the technical field of displacement calibration, and in particular to a machine vision displacement calibration method for foundation pit construction. Background Art

[0002] With the rapid development of urban construction, the construction of urban subways and super high-rise buildings is increasing, which brings with it the problem of foundation pit construction. The construction of foundation pit is a relatively risky project, so the calibration of foundation pit is particularly important.

[0003] However, traditional foundation pit displacement monitoring mainly relies on total stations, levels, etc., but the measurement accuracy of total stations and levels is not high, and they need to be guided by fixed reference points, the cumulative error is relatively large, and they are manual measurements and are also subject to errors caused by human influence. The level of automation is not high. At the same time, direct contact with the detection object is required. Long-term work causes instrument instability, reduces the accuracy of the results of judging the foundation pit displacement status, and cannot effectively guide the foundation pit excavation work.

[0004] Therefore, the present invention proposes a machine vision displacement calibration method for foundation pit construction. Summary of the Invention

[0005] The present invention provides a machine vision displacement calibration method for foundation pit construction, which is used to solve the defects of the existing technology that the measurement accuracy of total stations and levels is not high, and the points need to be guided through fixed reference points, the cumulative error is relatively large, and the manual measurement is also subject to errors caused by human influence, the level of automation is not high, and direct contact with the detection object is required. Long-term work causes instrument instability, reduces the accuracy of the results of judging the displacement state of the foundation pit, and cannot effectively guide the excavation work of the foundation pit.

[0006] In one aspect, the present invention provides a machine vision displacement calibration method for foundation pit construction, comprising:

[0007] Step 1: Obtain the structure and location of the target foundation pit construction, and obtain the corresponding camera according to the structure and location;

[0008] Step 2: Obtain the data of the camera's distortion parameters and internal focus parameters, and calibrate the camera according to the calibration algorithm;

[0009] Step 3: Preprocessing the image data of the foundation pit construction captured by the calibrated camera;

[0010] Step 4: extracting significant feature points of the foundation pit edge based on the preprocessed image data using a computer vision algorithm, predicting the significant feature points based on a support vector machine, and determining the actual location of the foundation pit based on the prediction results;

[0011] Step 5: Compare the actual position of the foundation pit with the original set position to obtain the displacement deviation of the foundation pit position, perform displacement calibration, adjust the driving trajectory of the excavator, and achieve accurate correction of the displacement of the foundation pit position.

[0012] According to the present invention, a machine vision displacement calibration method for foundation pit construction provided by the present invention further includes: before obtaining the structure and position of the target foundation pit construction and obtaining the corresponding camera according to the structure and position;

[0013] Obtain the excavation parameters of the target foundation pit, and determine the deformation parameters of the target foundation pit and the surrounding surface settlement parameters based on the energy conservation effect according to the excavation parameters;

[0014] Determine the benchmark visual difference of the target foundation pit horizontal displacement based on the deformation parameters and surrounding surface settlement parameters;

[0015] Deploy the benchmark reference target and monitoring reference target of the target foundation pit based on the benchmark visual difference, and determine the overlapping visual features of the benchmark reference target and the monitoring reference target in the three-dimensional coordinate system;

[0016] Determine the displacement error vector between the benchmark reference target and the monitoring reference target based on overlapping visual features based on the difference technology;

[0017] Determine the optimal horizontal displacement monitoring point position according to the displacement error vector, and obtain the terrain factor of the area where the optimal horizontal position monitoring point is located;

[0018] Determine the soil structure within the range of the optimal horizontal position monitoring point based on topographic factors, and determine the soil retention and soil loss characteristics within the range based on the soil structure;

[0019] Determine the critical value of the static balance maintenance factor and dynamic imbalance inducing factor of the optimal horizontal position monitoring point based on soil conservation characteristics and soil loss characteristics;

[0020] The periodic stability coefficient of the optimal horizontal position monitoring point is determined according to the critical point value, and whether the optimal horizontal position monitoring point needs to be reinforced is determined according to the periodic stability coefficient. If so, the optimal horizontal position monitoring point is reinforced and used as the installation position of the camera.

[0021] According to a machine vision displacement calibration method for foundation pit construction provided by the present invention, the structure and position of the target foundation pit construction are obtained, and the corresponding camera is obtained according to the structure and position, including:

[0022] Obtaining a three-dimensional model of the foundation pit construction using a laser scanner, and obtaining the structure and position of the target foundation pit construction based on the three-dimensional model;

[0023] Obtaining the structural type and structural characteristics of the target foundation pit construction;

[0024] The corresponding camera is acquired based on the resolution and clarity according to the structure type and structural characteristics and the size information of the foundation pit.

[0025] According to a machine vision displacement calibration method for foundation pit construction provided by the present invention, data on distortion parameters and internal focus parameters of a camera are obtained, including:

[0026] Use the camera calibration tool to calibrate the camera and obtain the camera's intrinsic parameters and distortion model based on the calibration results;

[0027] Get images taken by the camera at different angles and distances;

[0028] A plurality of influencing factors of the camera are obtained, an image is processed based on an image processing algorithm according to the plurality of influencing factors, and data of distortion parameters and internal focus parameters of the camera are obtained according to the processing results.

[0029] According to a machine vision displacement calibration method for foundation pit construction provided by the present invention, a camera is calibrated according to a calibration algorithm, comprising:

[0030] Match the camera's internal parameters to the actual geometric structure through rotation and translation, and adjust the camera's posture based on the matching results and visual feedback;

[0031] The camera after adjusting its posture is calibrated based on the known points and edges using a calibration algorithm, and the internal and external parameters of the camera are estimated based on the calibration results;

[0032] Performing tests based on different types of images according to the internal and external parameters of the camera, and verifying the accuracy of the internal and external parameters of the camera according to the test results;

[0033] The coordinates of a point are converted between the camera coordinate system and the world coordinate system through rotation and translation, including:

[0034] Set the midpoint of the camera coordinate system , after conversion to the world coordinate system , the relationship between the two is:

[0035]

[0036] Through the imaging model of the camera, the three-dimensional point in the camera coordinate system can be mapped to the two-dimensional point on the imaging plane. According to the pinhole imaging model, we can get , so the point on the camera coordinate system is Can be compared with points on the two-dimensional imaging plane Conversion between each other, the conversion formula is:

[0037]

[0038] Because the processing object is pixels during image processing, the pixel coordinates on the image coordinate system are With camera coordinates The relationship is:

[0039]

[0040] in is the focal point between the image plane and the camera optical axis, , is the ratio of the pixel to the length in the x and y directions, so the expression of the final imaging model can be obtained, and the relationship is

[0041]

[0042] Therefore, the change in the coordinates of the pixel point in the image coordinate system can be used to infer the change in the coordinates of the world coordinate system, and the difference between the coordinates of the pixel point on the initial imaging plane and the coordinates of the pixel point on the current imaging plane can be used to infer the difference between the current world coordinates and the initial world coordinates, and the actual displacement can be obtained.

[0043] According to a machine vision displacement calibration method for foundation pit construction provided by the present invention, image data of foundation pit construction captured by a calibrated camera is preprocessed, comprising:

[0044] A certain number of control points are arranged around the foundation pit, and image data of the foundation pit construction is captured using a calibrated camera based on the coverage range according to the control points;

[0045] The data is normalized and gray-scaled, and the processed image data is obtained and stored.

[0046] According to the present invention, a machine vision displacement calibration method for foundation pit construction is provided. The method extracts significant feature points of the foundation pit edge based on preprocessed image data using a computer vision algorithm, predicts the significant feature points based on a support vector machine, and determines the actual position of the foundation pit based on the prediction results, including:

[0047] Extract significant feature points on the edge of the foundation pit based on the preprocessed image data using computer vision algorithms;

[0048] Analyze the characteristic points to obtain specific locations of the characteristic points on the foundation pit boundary;

[0049] Acquire the distance and direction between the feature points according to the specific position;

[0050] Matching the extracted feature points with the feature points on the preprocessed image based on the distance and direction between the feature points based on a feature matching algorithm to obtain corresponding matching point pairs;

[0051] A camera pose estimation is obtained based on triangulated geometry according to the matching point pairs, prediction is performed based on a support vector machine according to the pose estimation, and the actual position of the foundation pit is determined according to the prediction result.

[0052] According to the present invention, a machine vision displacement calibration method for foundation pit construction is provided. The actual position of the foundation pit is compared with the original set position to obtain the displacement deviation of the foundation pit position, and the displacement is calibrated to adjust the driving trajectory of the excavator to achieve accurate correction of the foundation pit position displacement. The method includes:

[0053] The actual position of the foundation pit is imported into the measurement software through the data processing software for processing;

[0054] Obtain the displacement deviation between the actual position of the foundation pit and the original set position according to the displacement deviation formula;

[0055] The displacement deviation is judged, and whether the deviation is within a preset range is determined according to the judgment result. If not, the foundation pit is calibrated according to the reflector calibration method;

[0056] The excavator's driving trajectory is adjusted according to the calibration results to achieve accurate correction of the foundation pit position displacement.

[0057] Compared with the prior art, the present invention has the following advantages:

[0058] By calibrating the corresponding camera, image data of foundation pit construction is obtained and the actual position of the foundation pit is predicted based on the support vector machine. The predicted result is compared with the original set position to obtain the position offset of the foundation pit, which can improve the measurement accuracy. No manual measurement is required. At the same time, there is no need to directly contact the detection object, avoiding long-term work that may cause instrument instability, improving the accuracy of the results of judging the offset status of the foundation pit, and can effectively guide the excavation work of the foundation pit. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0060] Figure 1 This is a flow chart of a machine vision displacement calibration method for foundation pit construction provided by an embodiment of the present invention;

[0061] Figure 2 This is a flow chart of obtaining corresponding cameras according to the structure and position of the target foundation pit construction, provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0063] Example 1:

[0064] The embodiment of the present invention provides a machine vision displacement calibration method for foundation pit construction, such as Figure 1 As shown, the method mainly includes the following steps:

[0065] Step 1: Obtain the structure and location of the target foundation pit construction, and obtain the corresponding camera according to the structure and location;

[0066] Step 2: Obtain the data of the camera's distortion parameters and internal focus parameters, and calibrate the camera according to the calibration algorithm;

[0067] Step 3: Preprocessing the image data of the foundation pit construction captured by the calibrated camera;

[0068] Step 4: extracting significant feature points of the foundation pit edge based on the preprocessed image data using a computer vision algorithm, predicting the significant feature points based on a support vector machine, and determining the actual location of the foundation pit based on the prediction results;

[0069] Step 5: Compare the actual position of the foundation pit with the original set position to obtain the displacement deviation of the foundation pit position, perform displacement calibration, adjust the driving trajectory of the excavator, and achieve accurate correction of the displacement of the foundation pit position.

[0070] In this embodiment, the distortion parameters of the camera refer to a series of parameters that describe the image distortion or deformation caused by factors such as the lens or sensor during the camera imaging process, including:

[0071] Lens distortion parameters: describe the impact of the shape and size of the lens on imaging, such as pincushion distortion, barrel distortion, and vignetting distortion.

[0072] Sensor distortion parameters: describe the impact of the sensor's sensitivity to light and response characteristics on imaging, such as color drift, color distortion, and brightness drift.

[0073] In this embodiment, the computer vision algorithm is a machine learning method that simulates the functions of the human eye by using computer programs and algorithms, and is used to process video streams or image sequences from a camera and extract meaningful information therefrom.

[0074] The beneficial effects of the above technical solution are: by calibrating the corresponding camera, obtaining image data of foundation pit construction and predicting the actual position of the foundation pit based on the support vector machine, and comparing the predicted result with the original set position, the position offset of the foundation pit is obtained, which can improve the measurement accuracy without the need for manual measurement. At the same time, there is no need to directly contact the detection object, avoiding long-term work causing instrument instability, improving the accuracy of the results of judging the offset status of the foundation pit, and can effectively guide the excavation work of the foundation pit.

[0075] Example 2:

[0076] Based on Example 1, before obtaining the structure and position of the target foundation pit construction and obtaining the corresponding camera according to the structure and position, the embodiment of the present invention further includes:

[0077] Obtain the excavation parameters of the target foundation pit, and determine the deformation parameters of the target foundation pit and the surrounding surface settlement parameters based on the energy conservation effect according to the excavation parameters;

[0078] Determine the benchmark visual difference of the target foundation pit horizontal displacement based on the deformation parameters and surrounding surface settlement parameters;

[0079] Deploy the benchmark reference target and monitoring reference target of the target foundation pit based on the benchmark visual difference, and determine the overlapping visual features of the benchmark reference target and the monitoring reference target in the three-dimensional coordinate system;

[0080] Determine the displacement error vector between the benchmark reference target and the monitoring reference target based on overlapping visual features based on the difference technology;

[0081] Determine the optimal horizontal displacement monitoring point position according to the displacement error vector, and obtain the terrain factor of the area where the optimal horizontal position monitoring point is located;

[0082] Determine the soil structure within the range of the optimal horizontal position monitoring point based on topographic factors, and determine the soil retention and soil loss characteristics within the range based on the soil structure;

[0083] Determine the critical value of the static balance maintenance factor and dynamic imbalance inducing factor of the optimal horizontal position monitoring point based on soil conservation characteristics and soil loss characteristics;

[0084] The periodic stability coefficient of the optimal horizontal position monitoring point is determined according to the critical point value, and whether the optimal horizontal position monitoring point needs to be reinforced is determined according to the periodic stability coefficient. If so, the optimal horizontal position monitoring point is reinforced and used as the installation position of the camera.

[0085] In this embodiment, the excavation parameters of the target foundation pit generally refer to some important parameters that need to be considered when digging the foundation pit, such as: excavation depth, excavation area, excavation method, and support structure.

[0086] In this embodiment, the law of conservation of energy states that in a closed system, all energy can only be converted from one form to another, and the total amount remains constant.

[0087] In this embodiment, the deformation parameter of the target foundation pit generally refers to the degree of deformation of the foundation pit retaining structure under the action of external force.

[0088] In this embodiment, the surrounding surface settlement parameter refers to the degree of settlement of the surrounding surface of the foundation pit retaining structure when it is subjected to external forces.

[0089] In this embodiment, the optimal horizontal position monitoring point refers to a point selected by a specific measurement method in a certain scene, which is used to monitor the horizontal state of the target object to ensure that it is in a normal horizontal state.

[0090] The beneficial effect of the above technical solution is: the periodic stability coefficient of the optimal horizontal position monitoring point is judged according to the critical point value of the static balance maintenance factor and the dynamic imbalance inducing factor of the optimal horizontal position monitoring point, so as to determine whether the optimal horizontal position monitoring point needs to be reinforced, which can enable the camera to better resist the influence of external factors and help improve the quality and stability of the image.

[0091] Example 3:

[0092] Based on Example 2, the embodiment of the present invention obtains the structure and position of the target foundation pit construction, and obtains the corresponding camera according to the structure and position, such as Figure 2 Shown, including:

[0093] S01: Obtain a three-dimensional model of the foundation pit construction using a laser scanner, and obtain the structure and position of the target foundation pit construction according to the three-dimensional model;

[0094] S02: Obtaining the structural type and structural characteristics of the target foundation pit construction;

[0095] S03: Acquire a corresponding camera based on resolution and clarity according to the structure type and structural characteristics and size information of the foundation pit.

[0096] In this embodiment, the laser scanner is an instrument that uses laser technology to perform high-precision measurement on the surface of an object, and acquires three-dimensional data of the object by emitting a laser beam and measuring the returned reflected light signal.

[0097] In this embodiment, the three-dimensional model of foundation pit construction includes: groundwater level, geological conditions, and excavation depth.

[0098] The beneficial effects of the above technical solution are: by obtaining the corresponding camera according to the structure and position of the target foundation pit construction, the optimal camera installation position can be determined to ensure that the entire foundation pit construction site can be covered, so that all conditions on the site can be captured in real time, improving the real-time nature of monitoring. At the same time, by obtaining the corresponding target information, the camera parameters can be set according to its position, size and other characteristics, so that it can better capture changes in the target object, effectively avoid false alarms or missed alarms caused by environmental changes, and enhance the accuracy of monitoring.

[0099] Example 4:

[0100] Based on Example 3, the embodiment of the present invention obtains data on distortion parameters and internal focus parameters of the camera, including:

[0101] Use the camera calibration tool to calibrate the camera and obtain the camera's intrinsic parameters and distortion model based on the calibration results;

[0102] Get images taken by the camera at different angles and distances;

[0103] A plurality of influencing factors of the camera are obtained, an image is processed based on an image processing algorithm according to the plurality of influencing factors, and data of distortion parameters and internal focus parameters of the camera are obtained according to the processing results.

[0104] In this embodiment, the camera calibration tool may be: a calibration plate or a checkerboard.

[0105] In this embodiment, the intrinsic parameters of the camera refer to the physical characteristics of the camera itself, including focal length, number of pixels, principal point coordinates, and distortion coefficient.

[0106] In this embodiment, the distortion model is used to estimate the geometric deformation caused by the camera being misaligned, which may cause points in the image to be distorted.

[0107] In this embodiment, the multiple influencing factors of the camera include: principal point position, lens type, and aberration.

[0108] In this embodiment, the image processing algorithm may be: corner detection based on edge detection and triangulation.

[0109] The beneficial effects of the above technical solution are: obtaining the internal parameters and distortion model of the camera according to the correction results, and processing the image based on multiple influencing factors of the camera, obtaining the data of the distortion parameters and internal focus parameters of the camera, which can control the camera and effectively reduce the distortion, deformation and other problems of the foundation pit construction image, thereby improving the clarity and quality of the foundation pit construction image.

[0110] Example 5:

[0111] Based on Example 4, this embodiment of the present invention calibrates the camera according to the calibration algorithm, including:

[0112] Match the camera's internal parameters to the actual geometric structure through rotation and translation, and adjust the camera's posture based on the matching results and visual feedback;

[0113] The camera after adjusting its posture is calibrated based on the known points and edges using a calibration algorithm, and the internal and external parameters of the camera are estimated based on the calibration results;

[0114] Performing tests based on different types of images according to the internal and external parameters of the camera, and verifying the accuracy of the internal and external parameters of the camera according to the test results;

[0115] The coordinates of a point are converted between the camera coordinate system and the world coordinate system through rotation and translation, including:

[0116] Set the midpoint of the camera coordinate system , after conversion to the world coordinate system , the relationship between the two is:

[0117]

[0118] Through the imaging model of the camera, the three-dimensional point in the camera coordinate system can be mapped to the two-dimensional point on the imaging plane. According to the pinhole imaging model, we can get , so the point on the camera coordinate system is Can be compared with points on the two-dimensional imaging plane Conversion between each other, the conversion formula is:

[0119]

[0120] Because the processing object is pixels during image processing, the pixel coordinates on the image coordinate system are With camera coordinates The relationship is:

[0121]

[0122] in is the focal point between the image plane and the camera optical axis, , is the ratio of the pixel to the length in the x and y directions, so the expression of the final imaging model can be obtained, and the relationship is

[0123]

[0124] Therefore, the change in the coordinates of the pixel point in the image coordinate system can be used to infer the change in the coordinates of the world coordinate system, and the difference between the coordinates of the pixel point on the initial imaging plane and the coordinates of the pixel point on the current imaging plane can be used to infer the difference between the current world coordinates and the initial world coordinates, and the actual displacement can be obtained.

[0125] In this embodiment, the internal parameters of the camera refer to inherent parameters inside the camera, and are important factors for describing the imaging capability of the camera.

[0126] In this embodiment, matching the internal parameters of the camera with the actual geometric structure means that the parameters such as focal length, number of pixels, sensitivity, aperture set inside the camera are consistent with the actual geometric structure of the camera (such as lens type, sensor size, etc.), so as to ensure that the imaging quality and performance of the camera are in the best state.

[0127] In this embodiment, the calibration algorithm includes: a circular target ring and a polygon.

[0128] In this embodiment, different types of images include: planar images, stereoscopic images, and depth images.

[0129] The beneficial effects of the above technical solution are: by matching the internal parameters of the camera with the actual geometric structure, the clarity and quality of the captured image can be improved, and unnecessary noise and distortion can be reduced. At the same time, by adjusting the camera posture based on visual feedback, calibrating known points and edges based on the calibration algorithm, and estimating the internal and external parameters of the camera, the position and direction of the camera can be better calibrated and adjusted, thereby improving the accuracy and reliability of video surveillance.

[0130] Example 6:

[0131] Based on Example 5, this embodiment of the present invention pre-processes the image data of the foundation pit construction captured by the calibrated camera, including:

[0132] A certain number of control points are arranged around the foundation pit, and image data of the foundation pit construction is captured using a calibrated camera based on the coverage range according to the control points;

[0133] The data is normalized and gray-scaled, and the processed image data is obtained and stored.

[0134] In this embodiment, the control point may be a known reference point or a known coordinate.

[0135] In this embodiment, normalization is a data processing method used to convert a set of data into values ​​within the same scale or range to facilitate comparison and analysis.

[0136] In this embodiment, grayscale processing is an image processing method for converting a color image into a grayscale image and converting it into the same grayscale range. The grayscale processing is to eliminate color information to make the image more concise and clear.

[0137] The beneficial effect of the above technical solution is: based on the image data of the foundation pit construction captured by the calibrated camera, the image data is preprocessed, which can improve the quality of the image data and facilitate the subsequent extraction of feature points.

[0138] Example 7:

[0139] Based on Example 6, this embodiment of the present invention extracts significant feature points at the edge of a foundation pit based on preprocessed image data using a computer vision algorithm, predicts the significant feature points based on a support vector machine, and determines the actual position of the foundation pit based on the prediction results, including:

[0140] Extract significant feature points on the edge of the foundation pit based on the preprocessed image data using computer vision algorithms;

[0141] Analyze the characteristic points to obtain specific locations of the characteristic points on the foundation pit boundary;

[0142] Acquire the distance and direction between the feature points according to the specific position;

[0143] Matching the extracted feature points with the feature points on the preprocessed image based on the distance and direction between the feature points based on a feature matching algorithm to obtain corresponding matching point pairs;

[0144] A camera pose estimation is obtained based on triangulated geometry according to the matching point pairs, prediction is performed based on a support vector machine according to the pose estimation, and the actual position of the foundation pit is determined according to the prediction result.

[0145] In this embodiment, the significant feature points of the foundation pit edge refer to the ground boundary lines or ground feature contour lines that require special attention in the foundation pit project, that is, the locations that need to be identified and measured before construction.

[0146] In this embodiment, the feature matching algorithm detects and confirms the correspondence between two or more images in an image or video sequence by comparing local features between them.

[0147] The beneficial effects of the above technical solution are: by analyzing the significant feature points at the edge of the foundation pit, obtaining the specific position of the feature points on the boundary of the foundation pit, and determining the matching point pairs of the feature points on the preprocessed image, the accuracy of the matching points can be improved, laying the foundation for image registration. At the same time, prediction based on the support vector machine can detect dangerous points in advance, carry out prevention and control in advance, and improve the prediction results.

[0148] Example 8:

[0149] Based on Example 7, this embodiment of the present invention compares the actual position of the foundation pit with the original set position, obtains the displacement deviation of the foundation pit position, performs displacement calibration, adjusts the driving trajectory of the excavator, and realizes accurate correction of the displacement of the foundation pit position, including:

[0150] The actual position of the foundation pit is imported into the measurement software through the data processing software for processing;

[0151] Obtain the displacement deviation between the actual position of the foundation pit and the original set position according to the displacement deviation formula;

[0152] The displacement deviation is judged, and whether the deviation is within a preset range is determined according to the judgment result. If not, the foundation pit is calibrated according to the reflector calibration method;

[0153] The excavator's driving trajectory is adjusted according to the calibration results to achieve accurate correction of the foundation pit position displacement.

[0154] In this embodiment, the displacement deviation formula is: displacement deviation = (actual position - original position) / original position.

[0155] In this embodiment, the reflector calibration method is a method for calibrating the foundation pit by measuring the position of the image point of the camera in a ratio with the actual size under known conditions.

[0156] In this embodiment, adjusting the driving trajectory of the excavator refers to adjusting the driving direction and path of the excavator while the excavator is working.

[0157] The beneficial effect of the above technical solution is: by comparing the actual position of the foundation pit with the original set position and judging whether the deviation is within the preset range, the foundation pit is calibrated and the driving trajectory of the excavator is adjusted, the displacement of the foundation pit position can be accurately corrected, thereby avoiding construction errors and safety risks caused by foundation pit position deviation.

[0158] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0159] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A machine vision displacement calibration method for foundation pit construction, characterized in that: include: Step 1: Obtain the structure and location of the target foundation pit construction, and obtain the corresponding camera according to the structure and location; Step 2: Obtain the data of the camera's distortion parameters and internal focus parameters, and calibrate the camera according to the calibration algorithm; Step 3: Preprocessing the image data of the foundation pit construction captured by the calibrated camera; Step 4: extracting significant feature points of the foundation pit edge based on the preprocessed image data using a computer vision algorithm, predicting the significant feature points based on a support vector machine, and determining the actual location of the foundation pit based on the prediction results; Step 5: Compare the actual position of the foundation pit with the original set position to obtain the displacement deviation of the foundation pit position, perform displacement calibration, adjust the driving trajectory of the excavator, and achieve accurate correction of the displacement of the foundation pit position; Before obtaining the structure and position of the target foundation pit construction and obtaining the corresponding camera according to the structure and position, the method further includes: Obtain the excavation parameters of the target foundation pit, and determine the deformation parameters of the target foundation pit and the surrounding surface settlement parameters based on the energy conservation effect according to the excavation parameters; Determine the benchmark visual difference of the target foundation pit horizontal displacement based on the deformation parameters and surrounding surface settlement parameters; Deploy the benchmark reference target and monitoring reference target of the target foundation pit based on the benchmark visual difference, and determine the overlapping visual features of the benchmark reference target and the monitoring reference target in the three-dimensional coordinate system; Determine the displacement error vector between the benchmark reference target and the monitoring reference target based on overlapping visual features based on the difference technology; Determine the optimal horizontal displacement monitoring point position according to the displacement error vector, and obtain the terrain factor of the area where the optimal horizontal position monitoring point is located; Determine the soil structure within the range of the optimal horizontal position monitoring point based on topographic factors, and determine the soil retention and soil loss characteristics within the range based on the soil structure; Determine the critical value of the static balance maintenance factor and dynamic imbalance inducing factor of the optimal horizontal position monitoring point based on soil conservation characteristics and soil loss characteristics; The periodic stability coefficient of the optimal horizontal position monitoring point is determined according to the critical point value, and whether the optimal horizontal position monitoring point needs to be reinforced is determined according to the periodic stability coefficient. If so, the optimal horizontal position monitoring point is reinforced and used as the installation position of the camera.

2. The machine vision displacement calibration method for foundation pit construction according to claim 1, characterized in that: Obtaining the structure and location of the target foundation pit construction, and obtaining the corresponding camera according to the structure and location, including: Obtaining a three-dimensional model of the foundation pit construction using a laser scanner, and obtaining the structure and position of the target foundation pit construction based on the three-dimensional model; Obtaining the structural type and structural characteristics of the target foundation pit construction; The corresponding camera is acquired based on the resolution and clarity according to the structure type and structural characteristics and the size information of the foundation pit.

3. The machine vision displacement calibration method for foundation pit construction according to claim 1, characterized in that: Obtain the camera's distortion parameters and internal focus parameters, including: Use the camera calibration tool to calibrate the camera and obtain the camera's intrinsic parameters and distortion model based on the calibration results; Get images taken by the camera at different angles and distances; A plurality of influencing factors of the camera are obtained, an image is processed based on an image processing algorithm according to the plurality of influencing factors, and data of distortion parameters and internal focus parameters of the camera are obtained according to the processing results.

4. The machine vision displacement calibration method for foundation pit construction according to claim 1, characterized in that: Calibrate the camera according to the calibration algorithm, including: Match the camera's internal parameters to the actual geometric structure through rotation and translation, and adjust the camera's posture based on the matching results and visual feedback; The camera after adjusting its posture is calibrated based on the known points and edges using a calibration algorithm, and the internal and external parameters of the camera are estimated based on the calibration results; Performing tests based on different types of images according to the internal and external parameters of the camera, and verifying the accuracy of the internal and external parameters of the camera according to the test results; The coordinates of a point are converted between the camera coordinate system and the world coordinate system through rotation and translation, including: Set the midpoint of the camera coordinate system , after conversion to the world coordinate system , the relationship between the two is: Through the imaging model of the camera, the three-dimensional point in the camera coordinate system can be mapped to the two-dimensional point on the imaging plane. According to the pinhole imaging model, we can get , so the point on the camera coordinate system is Can be compared with points on the two-dimensional imaging plane Mutual conversion, the conversion formula is: Because the processing object is pixels during image processing, the pixel coordinates on the image coordinate system are With camera coordinates The relationship is: in is the focal point between the image plane and the camera optical axis, , is the ratio of the pixel to the length in the x and y directions, so the expression of the final imaging model can be obtained, and the relationship is Therefore, the change in the coordinates of the pixel point in the image coordinate system can be used to infer the change in the coordinates of the world coordinate system, and the difference between the coordinates of the pixel point on the initial imaging plane and the coordinates of the pixel point on the current imaging plane can be used to infer the difference between the current world coordinates and the initial world coordinates, and the actual displacement can be obtained.

5. The machine vision displacement calibration method for foundation pit construction according to claim 1, characterized in that: The image data of the foundation pit construction captured by the calibrated camera is preprocessed, including: A certain number of control points are arranged around the foundation pit, and image data of the foundation pit construction is captured using a calibrated camera based on the coverage range according to the control points; The data is normalized and gray-scaled, and the processed image data is obtained and stored.

6. The machine vision displacement calibration method for foundation pit construction according to claim 1, characterized in that: Extracting significant feature points at the edge of the foundation pit based on the preprocessed image data using a computer vision algorithm, predicting the significant feature points based on a support vector machine, and determining the actual position of the foundation pit based on the prediction results, including: Extract significant feature points on the edge of the foundation pit based on the preprocessed image data using computer vision algorithms; Analyze the characteristic points to obtain specific locations of the characteristic points on the foundation pit boundary; Acquire the distance and direction between the feature points according to the specific position; Matching the extracted feature points with the feature points on the preprocessed image based on the distance and direction between the feature points based on a feature matching algorithm to obtain corresponding matching point pairs; A camera pose estimation is obtained based on triangulated geometry according to the matching point pairs, prediction is performed based on a support vector machine according to the pose estimation, and the actual position of the foundation pit is determined according to the prediction result.

7. The machine vision displacement calibration method for foundation pit construction according to claim 1, characterized in that: Comparing the actual position of the foundation pit with the original set position to obtain the displacement deviation of the foundation pit position, and performing displacement calibration to adjust the driving trajectory of the excavator to achieve accurate correction of the displacement of the foundation pit position, including: The actual position of the foundation pit is imported into the measurement software through the data processing software for processing; Obtain the displacement deviation between the actual position of the foundation pit and the original set position according to the displacement deviation formula; The displacement deviation is judged, and whether the deviation is within a preset range is determined according to the judgment result. If not, the foundation pit is calibrated according to the reflector calibration method; The excavator's driving trajectory is adjusted according to the calibration results to achieve accurate correction of the foundation pit position displacement.

Images