Monitoring Method, Device, Equipment and Storage Medium for Pile Foundation Driving

Abstract

The monitoring method, device, equipment and storage medium for pile sinking of pile foundations provided by the present invention obtain pile foundation sinking images through a collection device arranged at a preset position. The pile foundation sinking images include the pile foundation to be monitored, and a pile body scale is provided on the pile foundation to be monitored. Calculate the actual size of a unit pixel according to the pile body scale to determine the vertical pixel calibration value. Determine the pile foundation sinking parameters according to the preset marking points and the vertical pixel calibration value. Input the pile foundation sinking parameters into a bearing capacity prediction model to obtain the predicted bearing capacity of the pile foundation to be monitored. By using an image collection device arranged at a preset position, the present invention can obtain pile foundation sinking images in real time and perform pixel calibration in combination with the pile body scale, accurately calculate the key parameters in the process of pile foundation sinking, thereby providing a reliable basis for the prediction of the bearing capacity of pile foundations, greatly improving the monitoring accuracy and efficiency, and avoiding the errors that may be brought by manual measurement.

Description

Technical Field

[0001] The present invention relates to the technical field of pile foundation driving, and particularly relates to a monitoring method, device, equipment and storage medium for pile foundation driving. Background Art

[0002] As an important infrastructure in construction engineering, the quality of the pile foundation driving process directly affects the stability and safety of the project. At present, the monitoring of the quality of pile foundation driving mostly relies on manual detection methods, or traditional instrument equipment is used to measure parameters such as pile driving depth, penetration degree and rebound value. These traditional methods have many disadvantages, such as being greatly affected by human factors, having high measurement errors, and being cumbersome to operate, and being unable to monitor and feedback the dynamic changes in the pile driving process in real time. Therefore, there is an urgent need for an efficient and accurate pile foundation driving monitoring method to improve the monitoring efficiency and accuracy.

[0003] With the development of digital image processing technology and deep learning algorithms, the pile foundation driving monitoring method based on image acquisition and analysis has gradually become a research hotspot. Although traditional image acquisition technology can provide rich image data, in practical applications, how to accurately extract pile driving parameters from images and conduct predictive analysis in combination with the actual engineering situation is still a technical problem.

[0004] In summary, the problems existing in the prior art need to be solved urgently. Summary of the Invention

[0005] The present invention provides a monitoring method, device, equipment and storage medium for pile foundation driving to solve the defects in the prior art and realize the real-time monitoring of the pile foundation driving process.

[0006] The present invention provides a monitoring method for pile foundation driving, including:

[0007] Obtain a pile foundation driving image through a collection device arranged at a preset position, where the pile foundation driving image includes a pile to be monitored, and a pile body scale is provided on the pile to be monitored;

[0008] Calculate the actual size of a unit pixel according to the pile body scale to determine a vertical pixel calibration value;

[0009] Determine pile foundation driving parameters according to a preset marking point and the vertical pixel calibration value;

[0010] Input the pile foundation driving parameters into a bearing capacity prediction model to obtain the predicted bearing capacity of the pile to be monitored.

[0011] According to the monitoring method for pile foundation driving provided by the present invention, the step of calculating the actual size of a unit pixel according to the pile body scale to determine a vertical pixel calibration value specifically includes:

[0012] Extract the area containing the pile body scale in the pile foundation pile driving image;

[0013] Extract the pixel positions of any two scale lines in the pile body scale in the image;

[0014] Determine the vertical pixel calibration value according to the actual distance and pixel distance between the scale lines, and the pixel distance is determined by the pixel positions of each scale line in the image.

[0015] According to a monitoring method for pile foundation pile driving provided by the present invention, the pile foundation pile driving parameters include the hammer penetration degree, the hammer rebound value, and the penetration depth.

[0016] According to a monitoring method for pile foundation pile driving provided by the present invention, the preset marking points include tracking points. The step of determining the pile foundation pile driving parameters according to the preset marking points and the vertical pixel calibration value specifically includes:

[0017] Determine the vertical pixel coordinates of the current tracking point according to the current pile foundation pile driving image;

[0018] Determine the vertical pixel coordinates of the initial tracking point according to the initial pile foundation pile driving image;

[0019] Determine the pile body penetration displacement change curve according to the vertical pixel coordinates of each current tracking point and the vertical pixel coordinates of the initial tracking point. The pile body penetration displacement change curve is used to characterize the change curve of the pile body penetration displacement over time;

[0020] Determine the hammer penetration value and the hammer rebound value according to the pile body penetration displacement change curve;

[0021] Among them, the current tracking point and the initial tracking point are both set on the pile body scale;

[0022] Determine the initial penetration depth of the pile foundation to be monitored according to the scale value where the current tracking point is located, the absolute elevation of the pile body scale position where the tracking point is located, and the absolute elevation of the ground surface at the pile position;

[0023] Determine the pile body penetration depth value during the construction process according to the initial penetration depth and the pile body penetration displacement change curve.

[0024] According to a monitoring method for pile foundation pile driving provided by the present invention, before the step of determining the pile foundation pile driving parameters according to the preset marking points and the vertical pixel calibration value, the method further includes:

[0025] When it is monitored that the coordinates of the tracking point exceed the preset threshold Update the position of the tracking point according to the preset formula, and update the reference sub-region matrix according to the position of the tracking point.

[0026] The preset formula is as follows:

[0027]

[0028] Wherein, is the vertical pixel coordinate of the updated tracking point, is the preset update distance, is the pixel calibration value, is the vertical pixel coordinate of the current tracking point.

[0029] According to a monitoring method for pile foundation sinking provided by the present invention, after the step of obtaining the pile foundation sinking image through the acquisition device arranged at the preset position, the method further includes:

[0030] Preprocess the pile foundation sinking image, and the preprocessing includes denoising, grayscale conversion, and contrast enhancement.

[0031] According to a monitoring method for pile foundation sinking provided by the present invention, the bearing capacity prediction model is trained in the following manner:

[0032] Obtain historical pile foundation sinking parameters;

[0033] Input the historical pile foundation sinking parameters into the pre-constructed bearing capacity prediction model for training until the convergence function is satisfied;

[0034] The bearing capacity prediction model is constructed based on a multi-layer perceptron.

[0035] The present invention also provides a monitoring device for pile foundation sinking, including:

[0036] An image acquisition module, configured to obtain a pile foundation sinking image through an acquisition device arranged at a preset position, where the pile foundation sinking image includes a pile foundation to be monitored, and a pile body scale is provided on the pile foundation to be monitored;

[0037] A calibration determination module, configured to calculate the actual size of a unit pixel according to the pile body scale to determine the vertical pixel calibration value;

[0038] A parameter determination module, configured to determine pile foundation sinking parameters according to a preset marking point and the vertical pixel calibration value;

[0039] A bearing capacity prediction module, configured to input the pile foundation sinking parameters into a bearing capacity prediction model to obtain the predicted bearing capacity of the pile foundation to be monitored.

[0040] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where when the processor executes the program, it implements the monitoring method for pile foundation sinking as described in any one of the above.

[0041] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the monitoring method for pile foundation driving as described in any one of the above.

[0042] The present invention also provides a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the monitoring method for pile foundation driving as described in any one of the above.

[0043] The monitoring method, device, equipment and storage medium for pile foundation driving provided by the present invention obtain a pile foundation driving image through an acquisition device arranged at a preset position, where the pile foundation driving image includes a pile to be monitored, and a pile body scale is provided on the pile to be monitored; calculate the actual size of a unit pixel according to the pile body scale to determine a vertical pixel calibration value; determine pile foundation driving parameters according to a preset marking point and the vertical pixel calibration value; input the pile foundation driving parameters into a bearing capacity prediction model to obtain the predicted bearing capacity of the pile to be monitored. The monitoring method for pile foundation driving provided by the present invention can, by using an image acquisition device arranged at a preset position, obtain a pile foundation driving image in real time and perform pixel calibration in combination with a pile body scale, accurately calculate key parameters in the process of pile foundation driving, such as hammer penetration, hammer rebound value and penetration depth, etc., so as to provide a reliable basis for the bearing capacity prediction of the pile foundation. This method can obtain pile driving data in real time and non-destructively through digital image technology, greatly improving the monitoring accuracy and efficiency, and avoiding errors that may be brought by manual measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0045] Figure 1 is a flowchart of the monitoring method for pile foundation driving provided by the present invention;

[0046] Figure 2 is a structural schematic diagram of the monitoring device for pile foundation driving provided by the present invention;

[0047] Figure 3 is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Obviously, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0049] To solve the problems in the prior art, the present invention proposes a monitoring method for pile driving of pile foundations, which improves the monitoring accuracy and efficiency and avoids the errors that may be brought by manual measurement. The monitoring method for pile driving of pile foundations will be described below, as Figure 1 shown, including but not limited to the following steps:

[0050] Step 110: Obtain an image of pile driving of a pile foundation through an acquisition device arranged at a preset position. The image of pile driving of the pile foundation includes the pile foundation to be monitored, and a pile body scale is provided on the pile foundation to be monitored.

[0051] In this step, first, a suitable position needs to be selected to install the image acquisition device. Usually, this device can be a high-definition camera or a video camera. To ensure the high precision of the monitoring process, the device should have sufficient resolution and a high video frame rate (for example, ≥100 frames per second) to avoid data loss and improve the real-time performance of the monitoring. The device should be set at a fixed preset position to ensure that the image acquisition perspective can cover the whole process of pile driving of the pile foundation.

[0052] A pile body scale should be provided on the pile foundation to be monitored. Usually, the pile body scale is a visible scale line perpendicular to the pile body. The function of this scale is to calibrate the relationship between the pixels in the image and the actual size in the subsequent steps, so as to ensure the measurement accuracy. When acquiring the image, the image should clearly contain the pile foundation and its pile body scale for subsequent processing.

[0053] Step 120: Calculate the actual size of a unit pixel according to the pile body scale to determine the vertical pixel calibration value.

[0054] The core of this step is to calibrate the pile body scale through image processing technology to determine the actual size corresponding to a unit pixel in the image. First, extract the area containing the pile body scale in the image of pile driving of the pile foundation. Then, through an image analysis algorithm, identify the position of each scale line on the scale, specifically extract the position of the scale line through the pixel coordinates in the image.

[0055] Next, by calculating the actual distance between two adjacent scale lines on the scale and their corresponding pixel distances in the image, the actual size of a unit pixel can be obtained. For example, if the actual distance between adjacent scale lines is 1 meter, and the corresponding pixel distance in the image is 50 pixels, then the actual size represented by each pixel is 1 meter / 50 pixels, that is, 0.02 meters / pixel. Through this method, the vertical pixel calibration value, that is, the actual height value corresponding to a unit pixel, can be obtained.

[0056] Step 130: Determine the pile driving parameters of the pile foundation according to the preset marking points and the vertical pixel calibration value.

[0057] In this step, multiple preset marking points need to be set in the image first. Common marking points include tracking points and landmark points. The tracking points are generally selected as the corner points of specific scale lines on the pile body scale, and the landmark points are selected as fixed points on the ground to help calculate the penetration depth.

[0058] Then, identify the tracking points and landmark points in the image, and determine their pixel coordinates in the image through image processing technology. By comparing the images before and after hammering, obtain the vertical pixel coordinates of the tracking points and landmark points before hammering, after hammering, and at the initial and end states.

[0059] According to the vertical pixel calibration value, the pixel coordinates in the image can be converted into actual distances. For example, through the pixel coordinate difference between the tracking point and the landmark point before hammering, the penetration depth before hammering can be calculated. Similarly, the pixel coordinate difference between the tracking point and the landmark point after hammering is used to calculate the hammering rebound value. By comparing the pixel coordinate differences between the initial state and the end state, the penetration depth during the entire pile driving process can be calculated.

[0060] This step can further accurately determine the pile driving parameters of the pile foundation, such as the scale value of the tracking point on the scale, the initial absolute elevation value, the hammering penetration, the hammering rebound value, and the penetration depth, etc., thus providing basic data for subsequent bearing capacity prediction.

[0061] Step 140: Input the pile driving parameters of the pile foundation into the bearing capacity prediction model to obtain the predicted bearing capacity of the pile foundation to be monitored.

[0062] In this step, the pile driving parameters of the pile foundation obtained through the previous steps are input into a pre-trained bearing capacity prediction model. This prediction model can be a deep learning-based model, such as a multi-layer perceptron (MLP), which is trained through a large amount of historical pile driving data of pile foundations. During the training process, the model will learn the non-linear relationship between pile foundation parameters (such as pile driving depth, penetration, etc.) and the bearing capacity of the pile foundation.

[0063] The specific structure of the multi-layer perceptron (MLP) can be:

[0064] (1)Number of input layer nodes: 7 - 12, corresponding to different input parameter types;

[0065] (2)Number of hidden layers: 3 - 5;

[0066] (3)Activation function: ReLU function;

[0067] (4)Optimization algorithm: Error backpropagation algorithm.

[0068] After inputting the pile driving parameters, the bearing capacity prediction model will calculate the predicted bearing capacity of the pile foundation in real time. The output result of the model is the estimated value of the bearing capacity of the pile foundation, which can help engineers evaluate whether the quality of the current pile driving meets the design requirements, and then adjust or optimize the construction process. Moreover, when any of the following situations is detected, an early warning is triggered: (1) The penetration per single hammer blow exceeds the design threshold by ±30%; (2) The rebound values of three consecutive hammer blows exceed the set safety value; so that the staff can handle it in time.

[0069] Through this series of steps, it is possible to achieve precise monitoring of the pile driving process of the pile foundation and bearing capacity prediction, effectively improve the accuracy and real-time performance of pile driving construction, and ensure the quality and safety of the project.

[0070] As a further optional embodiment, the step of calculating the actual size of a unit pixel according to the pile body scale to determine the vertical pixel calibration value specifically includes:

[0071] Extract the area containing the pile body scale in the pile foundation pile driving image;

[0072] Extract the pixel positions of any two scale lines in the pile body scale in the image;

[0073] Determine the vertical pixel calibration value according to the actual distance and pixel distance between adjacent scale lines, and the pixel distance is determined by the pixel positions of each scale line in the image.

[0074] As a further optional embodiment, the present invention can also adopt the following technical solution to implement the step of calculating the actual size of a unit pixel according to the pile body scale to determine the vertical pixel calibration value.

[0075] First, in the obtained pile foundation pile driving image, perform an extraction operation on the image area. This image usually includes the pile driving process of the pile foundation and the pile body scale. In order to accurately calculate the vertical pixel calibration value, it is necessary to accurately extract the pile body scale in the image. Image processing algorithms can be used to identify and extract the area related to the pile body scale in the image.

[0076] When extracting the scale area, techniques such as image segmentation and edge detection can be used to accurately define the position of the scale area. The extracted image will only contain the pile body scale area, providing basic data for subsequent processing.

[0077] After the extraction of the scale area of the pile shaft is completed, the next step is to identify and extract any two scale lines in the pile shaft scale. These scale lines are usually arranged at equal intervals and appear as a series of horizontal or vertical straight lines in the image. Try to extract the scale lines with a distance exceeding a preset threshold to ensure accurate measurement.

[0078] Through image processing algorithms (such as edge detection, morphological transformation, etc.), the pixel coordinates of the scale lines can be identified and extracted. These pixel coordinates will provide key data for subsequent calculations. During the extraction process, special attention needs to be paid to the resolution of the scale lines and their precise positions in the image to ensure that the pixel positions of each scale line can be accurately obtained.

[0079] After the extraction of the scale lines of the pile shaft scale is completed, the vertical pixel calibration value is determined by comparing the actual distance between two adjacent scale lines and their pixel distance in the image.

[0080] Specifically, the actual distance between the scale lines is known and can usually be obtained according to the design drawing or the actual size of the scale. For example, assume that the actual distance between every two adjacent scale lines is 10 cm. Then, through image processing software or algorithms, the pixel position difference between adjacent scale lines in the image is obtained, and the pixel distance is calculated. For example, assume that the pixel distance between two adjacent scale lines in the image is 50 pixels. Subsequently, according to the actual distance and the pixel distance, the actual size corresponding to each pixel is determined. If the actual distance between the scale lines is 10 cm and the pixel distance is 50 pixels, then the vertical pixel calibration value is: 10 cm / 50 pixels = 0.2 cm / pixel. In this way, the vertical pixel calibration value is determined, providing a basis for the calculation of the pile driving process parameters.

[0081] As a further optional embodiment, the pile driving parameters of the pile foundation include the hammer penetration degree, the hammer rebound value, and the penetration depth.

[0082] As a further optional embodiment, the preset marking points include tracking points. The step of determining the pile driving parameters of the pile foundation according to the preset marking points and the vertical pixel calibration value specifically includes:

[0083] Determine the vertical pixel coordinates of the current tracking point according to the current pile driving image of the pile foundation;

[0084] Determine the vertical pixel coordinates of the initial tracking point according to the initial pile driving image of the pile foundation;

[0085] Determine the pile shaft penetration displacement change curve according to the vertical pixel coordinates of each current tracking point and the vertical pixel coordinates of the initial tracking point. The pile shaft penetration displacement change curve is used to characterize the change curve of the pile shaft penetration displacement over time;

[0086] Determine the hammer penetration value and the hammer rebound value according to the pile body penetration displacement change curve;

[0087] Among them, both the current tracking point and the initial tracking point are set on the pile body scale;

[0088] Determine the initial penetration depth of the pile foundation to be monitored according to the scale value where the current tracking point is located, the absolute elevation of the pile body scale position where the tracking point is located, and the absolute elevation of the ground surface at the pile position;

[0089] Determine the pile body penetration depth value during the construction process according to the initial penetration depth and the pile body penetration displacement change curve.

[0090] First, by using the vertical pixel coordinates of the tracking point in the preset marking points in each pile foundation sinking image and combining with the vertical pixel calibration value (i.e., the relationship between pixels and actual dimensions), the displacement of the pile body at different time points can be calculated. Through these displacement data, a pile body penetration displacement change curve can be drawn, which describes the change of displacement with time during the sinking process of the pile foundation.

[0091] The specific steps are as follows:

[0092] Take the vertical pixel coordinates of the tracking point in each frame of the image T tyi (i.e., the vertical coordinate of the tracking point in the current frame of the image), and according to the aforementioned formula, combined with the vertical pixel calibration value cp , calculate the displacement:

[0093]

[0094] Among them R ty is the vertical coordinate of the tracking point in the initial frame of the image (i.e., the 0th frame of the image).

[0095] Through the displacement data of all frames of the image, the pile body penetration displacement change curve during the pile foundation sinking process can be obtained. This curve can reflect the change of the sinking depth of the pile foundation at each time point.

[0096] Step 2: Determine the hammer penetration value and the hammer rebound value

[0097] According to the obtained pile body penetration displacement change curve, the hammer penetration value and the hammer rebound value of each hammer can be further determined:

[0098] Hammer penetration value: When the pile foundation sinks with each hammer, the part of the displacement curve representing sinking is the penetration degree. The part of the displacement curve from zero to the maximum value is the penetration during the hammering process.

[0099] Hammer rebound value: After the hammering is completed, the pile foundation has a certain rebound. The part where the displacement curve shows a negative value or rebounds after hammering is the rebound value.

[0100] On the displacement change curve, a positive value indicates penetration, and a negative value indicates rebound. In this way, the penetration degree and rebound value of each hammer can be accurately obtained.

[0101] Step 3: Determine the vertical pixel coordinate of the current tracking point

[0102] Based on the current pile foundation sinking image, identify and extract the vertical pixel coordinate of the tracking point T txy At this time, the vertical coordinate of the tracking point T tyi represents the vertical displacement of the pile body position in the current image frame.

[0103] Step 4: Determine the vertical pixel coordinate of the initial tracking point

[0104] Based on the initial pile foundation sinking image (i.e., the 0th frame image), determine the vertical pixel coordinate of the initial tracking point T 0ty , as the reference coordinate. This coordinate represents the pile body position of the pile foundation to be monitored at time 0.

[0105] Step 5: Determine the penetration depth

[0106] Combining the vertical pixel coordinate of the current tracking point T txy 、 the vertical pixel coordinate of the initial tracking point T 0ty , the absolute construction elevation e1 of the initial tracking point and the scale value ht of the scale corresponding to the pile body where it is located, and the absolute construction elevation e0 at the pile position, the penetration depth of the pile foundation can be calculated. Specifically, the calculation formula for the penetration depth is as follows:

[0107]

[0108] Among them:

[0109] h t is the scale value corresponding to the tracking point on the pile body scale, indicating the height position of the tracking point on the pile body.

[0110] cp is the vertical pixel calibration value.

[0111] Through the above steps, the penetration depth of the pile foundation can be obtained in real time at different time points, and further analyze the bearing capacity and construction quality of the pile foundation during the sinking process.

[0112] As a preferred embodiment, for the curve of pile penetration displacement, the dynamic matching of tracking points can be realized based on the digital image correlation algorithm, including: constructing a reference sub-region matrix and a search region matrix, realizing image feature matching through two-dimensional Fourier transform, and adopting a sub-pixel positioning algorithm to improve the matching accuracy. Specifically:

[0113] Regard the image frame with reference points set as the reference image, denoted as the 0th frame (F0), and regard the subsequent ith frame as the target image, denoted as Fi. Use the digital image correlation algorithm to find the positions of the tracking points and landmark points in the target image, denoted as Ti ( , ), and Gi ( , ), respectively. Taking the matching algorithm of the tracking points as an example, the specific method is as follows:

[0114] 1) Take a square matrix with a side length of (2M + 1) centered at point T0 in the pixel matrix of F0 as the reference sub-region, denoted as R;

[0115] 2) Considering that the pile mainly moves vertically, take a matrix with a height of (6M + 1) and a width of (4M + 1) centered at the pixel coordinates ( , ) of point T(i - 1) in the pixel matrix of Fi as the search region, denoted as S, that is, S is the pixel matrix within the range from the -2Mth row to the +4Mth row and from the -2Mth column to the +2Mth column in the ith frame image;

[0116] 3) Insert M rows and M columns of 0 elements on each of the upper, lower, left, and right sides of the S matrix to expand it into a matrix of (8M + 1) rows and (6M + 1) rows, and calculate its two-dimensional Fourier transform, denoted as S_FFT (complex matrix);

[0117] 4) Flip the square matrix R by 180 degrees, insert 3M rows of 0 elements above and below it, and insert 2M rows of 0 elements on each of its left and right sides to expand it into a matrix of (8M + 1) rows and (6M + 1) rows, and calculate its two-dimensional Fourier transform, denoted as R_FFT (complex matrix);

[0118] 5) Perform element-wise multiplication (i.e., multiply each element, and the products of elements in the same position form a new matrix) on the matrices of S__FFT and R_FFT, and perform the inverse Fourier transform on the result, and then centralize the matrix obtained after the inverse transform (i.e., divide the matrix into four quadrants with the center as the origin, and exchange the elements in the first quadrant and the third quadrant, and exchange the elements in the second quadrant and the fourth quadrant), and the result is denoted as matrix C (complex matrix);

[0119] 6) In matrix C, take a matrix with (M, M) as the starting point, a height of (6M + 1), and a width of (4M + 1), and only retain the real part of all elements, denoted as Z_C;

[0120] 7) Locate the position of the maximum value in the square matrix Z_C, denoted as ( , ). This position is the integer pixel position in the search area S of Fi that has the highest correlation with the reference sub-region R;

[0121] 8) To further improve the monitoring accuracy, a local quadratic surface fitting algorithm can be used to more accurately calculate the position of the peak point. The specific method is as follows:

[0122] ① In the square matrix Z_C, take a square matrix with a side length of 3 centered at ( , ), and reduce its dimension by columns to a column vector, denoted as U;

[0123] ② Calculate the fitting coefficient A by cross-multiplying matrix X and U:

[0124] Where:

[0125]

[0126] ③ Calculate the fitting offset (sub-pixel)

[0127]

[0128]

[0129] Where:

[0130]

[0131] ④ Calculate the sub-pixel peak point position ( , ):

[0132]

[0133] 9) The position of the correlation coefficient peak point calculated in step 8) is the local sub-pixel coordinate position of the tracking point in the square matrix of the search area S in Fi. The pixel coordinate value of the tracking point in the global pixel matrix (Fi) can be calculated through the following formula, that is, ( , ):

[0134]

[0135]

[0136] As a further optional embodiment, before the step of determining the pile driving parameters of the pile foundation according to the preset marking points and the vertical pixel calibration values, the method further includes:

[0137] When it is monitored that the coordinates of the tracking point exceed the preset threshold the position of the tracking point is updated according to a preset formula, and the reference sub-region matrix is updated according to the position of the tracking point;

[0138] The preset formula is as follows:

[0139]

[0140] where is the longitudinal pixel coordinate of the updated tracking point, is the preset update distance, is the pixel calibration value, is the vertical pixel coordinate of the current tracking point.

[0141] In the actual monitoring process of pile driving of pile foundation, due to the influence of environmental factors or image quality, the tracking points in the pile driving images of pile foundation may not be clearly recognizable. For example, the tracking points may not be accurately located due to reasons such as shadows, occlusions, blurs, or image noises. At this time, in order to ensure the accurate calculation of the pile driving parameters, the system needs to automatically update the positions of the tracking points.

[0142] Specifically, first, the system performs a quality assessment on the pile driving images of pile foundation to detect whether there is a situation where the tracking points cannot be recognized. It can be judged whether the tracking points are clear through image processing techniques (such as edge detection, brightness contrast analysis, etc.).

[0143] When the system detects that the tracking points cannot be recognized, that is, when it is monitored that the coordinates of the tracking points exceed the preset threshold an update mechanism is triggered. In the pile driving images of pile foundation, the system will automatically select a suitable position on the pile body scale and reset the positions of the tracking points. Usually, a more obvious and unoccluded scale line position on the pile body scale can be selected as the new tracking point.

[0144] Once the new positions of the tracking points are selected, the system will re-determine the pixel coordinates of the new tracking points in the image through image processing techniques, such as image registration or feature matching algorithms. The new tracking points will be used for subsequent calculation of the pile driving parameters.

[0145] To ensure the accurate position of the updated tracking points, the vertical pixel calibration values (such as the calibration process in the previous steps) can be combined to accurately locate the updated tracking points and ensure their one-to-one correspondence with the actual positions on the pile body scale.

[0146] After the tracking points are updated, continue to perform the step of determining the pile driving parameters of the pile foundation according to the preset marker points and the vertical pixel calibration values. At this time, the updated tracking points will participate in the calculation of the pile driving parameters as one of the key marker points to ensure the accuracy and reliability of the monitoring data throughout the pile driving process.

[0147] As a further optional embodiment, after the step of obtaining the pile driving image of the pile foundation by the acquisition device arranged at the preset position, the method further includes:

[0148] Preprocess the pile driving image of the pile foundation, and the preprocessing includes denoising, grayscale conversion, and contrast enhancement.

[0149] The purpose of the preprocessing is to improve the quality of the pile driving image of the pile foundation, make the key features in the image more obvious, and facilitate subsequent parameter calculation and feature extraction.

[0150] Denoising processing: Noise in the image is often an important factor affecting the image quality. The noise may come from the shooting device, environmental light changes, or image transmission, etc. In the denoising processing of the pile driving image of the pile foundation, appropriate denoising algorithms, such as Gaussian filtering, median filtering, etc., are used to remove the random noise in the image. This can improve the clarity of the image and avoid noise interfering with subsequent processing.

[0151] Grayscale conversion: The color information in the pile driving image of the pile foundation has little influence on the extraction of the pile driving parameters, while grayscale images have higher computational efficiency in subsequent tasks such as edge detection and feature extraction. Therefore, the color image can be converted into a grayscale image.

[0152] Contrast enhancement: The contrast of the image determines the brightness difference between different regions in the image, and high contrast helps to better identify the features in the image (such as the edges of the pile foundation and the scales of the ruler). Through contrast enhancement, the details of the image can be made clearer and the distinction between the target features and the background can be enhanced.

[0153] As a further optional embodiment, the bearing capacity prediction model is trained in the following manner:

[0154] Obtain the historical pile driving parameters of the pile foundation;

[0155] Input the historical pile driving parameters of the pile foundation into the pre-constructed bearing capacity prediction model for training until the convergence function is satisfied;

[0156] The bearing capacity prediction model is constructed based on a multi-layer perceptron.

[0157] Before training the bearing capacity prediction model, it is first necessary to collect and organize historical pile driving parameter data of pile foundations. Historical data refers to the known pile driving parameters of pile foundations and the corresponding pile foundation bearing capacity data during past construction and monitoring processes. This data can be obtained through traditional testing methods, existing equipment, or established pile foundation monitoring systems.

[0158] Historical pile driving parameters of pile foundations generally include but are not limited to the following:

[0159] Penetration under hammering: The vertical settlement of the pile body when the pile foundation is hammered.

[0160] Rebound value under hammering: The degree of rebound of the pile foundation after being hammered, reflecting the elasticity of the pile foundation.

[0161] Penetration depth: The depth at which the pile foundation is inserted into the soil, which is usually related to the settlement of the pile foundation and the soil bearing capacity.

[0162] Soil layer characteristics: including soil density, viscosity, particle size, etc.

[0163] When predicting the bearing capacity, in addition to the penetration, rebound value, and penetration depth, static parameters such as hammer weight (i.e., the weight of the hammer), drop height (i.e., the height of the falling hammer), pile type (PHC pile, steel pipe pile), and pile diameter (the size of the pile) are also required.

[0164] When performing pile driving operations for pile foundations, in addition to real-time monitoring of dynamic parameters during the pile driving process, the following static parameters also need to be recorded:

[0165] Hammer weight: Refers to the weight of the hammer used to strike the pile body. The hammer weight has a direct impact on the settlement speed and driving depth of the pile foundation. Therefore, it is an important factor in predicting the bearing capacity of the pile foundation.

[0166] Drop height: Also known as the height of the falling hammer, it refers to the height at which the hammer freely falls from a high position. The drop height is related to dynamic parameters such as the settlement and rebound value of the pile foundation and is another key factor affecting the bearing capacity of the pile foundation.

[0167] Pile type: Different types of pile foundations (such as PHC piles, steel pipe piles, concrete piles, etc.) have differences in their materials, structures, and bearing characteristics. Therefore, the pile type is an important static parameter affecting the bearing capacity prediction.

[0168] Pile diameter: The diameter or size of the pile directly affects the bearing area of the pile foundation and thus its bearing capacity. Pile foundations with different pile diameters have different bearing capacity characteristics.

[0169] By collecting these historical pile driving parameters of pile foundations and combining them with the corresponding bearing capacity measurement data, the input-output pair data of the pile foundation is obtained.

[0170] Once the historical data is prepared, these pile driving parameters of the piles are used as inputs, together with the known pile bearing capacity as the output, and input into the pre-constructed bearing capacity prediction model. This prediction model can be a multi-layer perceptron (MLP) model based on artificial neural network (ANN).

[0171] The basic steps of the training process are as follows:

[0172] Data normalization: Normalize or standardize the input parameters to ensure that each feature contributes evenly to the model training and avoid biases between features with different dimensions. Common methods include min-max normalization, Z-score standardization, etc.

[0173] Model initialization: In a neural network, it is very important to initialize the network parameters (such as weights and biases). Initialization methods can use random initialization or other more advanced techniques (such as Xavier initialization, He initialization) to ensure that the network is in a reasonable state at the start of training.

[0174] Forward propagation: Input the normalized pile driving parameters of the piles into the multi-layer perceptron model and perform forward propagation calculations to obtain the predicted bearing capacity output.

[0175] Error calculation: Calculate the error between the predicted value and the actual value. Common error functions include mean squared error (MSE), etc.

[0176] Backward propagation and parameter update: Adjust the weights and biases in the network according to the error through the backpropagation algorithm. Usually, gradient descent optimization algorithms are used, such as stochastic gradient descent (SGD), Adam optimizer, etc.

[0177] Number of training epochs: Repeat the processes of forward propagation, error calculation, backward propagation, and parameter update until a predetermined stopping condition is met, such as reaching a certain number of training epochs or the error converges to a small enough value.

[0178] During the training process, the model will continuously update the parameters until the loss function reaches the preset convergence criterion. Common convergence criteria include:

[0179] Convergence of the loss function: When the change in the loss function is less than a certain threshold, it indicates that the model has reached the optimal or approximately optimal state.

[0180] Number of training epochs: Set the maximum number of training epochs, and stop training when the training reaches the predetermined number of times.

[0181] Performance on the validation set: Use the validation set to evaluate the generalization ability of the model, and stop training when the performance on the validation set no longer improves.

[0182] In this way, the bearing capacity prediction model is gradually optimized and can accurately predict the bearing capacity of the pile foundation from the input pile driving parameters.

[0183] In this embodiment, the bearing capacity prediction model is constructed based on a multi-layer perceptron (MLP). The MLP is a feed-forward neural network that contains multiple layers of neurons. The neurons in each layer take the output of the previous layer as input and finally output the predicted bearing capacity of the pile foundation.

[0184] The monitoring device for pile driving of pile foundation provided by the present invention will be described below. As Figure 2 shown, the monitoring device for pile driving of pile foundation described below can be correspondingly referred to the monitoring method for pile driving of pile foundation described above.

[0185] A monitoring device for pile driving of pile foundation includes:

[0186] An image acquisition module 210, configured to obtain an image of pile driving of pile foundation through an acquisition device arranged at a preset position. The image of pile driving of pile foundation includes the pile foundation to be monitored, and a pile body scale is provided on the pile foundation to be monitored;

[0187] A calibration determination module 220, configured to calculate the actual size of a unit pixel according to the pile body scale to determine a vertical pixel calibration value;

[0188] A parameter determination module 230, configured to determine pile driving parameters of the pile foundation according to preset marking points and the vertical pixel calibration value;

[0189] A bearing capacity prediction module 240, configured to input the pile driving parameters of the pile foundation into the bearing capacity prediction model to obtain the predicted bearing capacity of the pile foundation to be monitored.

[0190] Figure 3 The schematic physical structure diagram of an electronic device is exemplified. As Figure 3 shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340. Among them, the processor 310, the communication interface 320, and the memory 330 complete mutual communication through the communication bus 340. The processor 310 can call the logical instructions in the memory 330 to execute the monitoring method for pile driving of pile foundation, and the method includes:

[0191] Obtain an image of pile driving of pile foundation through an acquisition device arranged at a preset position. The image of pile driving of pile foundation includes the pile foundation to be monitored, and a pile body scale is provided on the pile foundation to be monitored;

[0192] Calculate the actual size of a unit pixel according to the pile body scale to determine a vertical pixel calibration value;

[0193] Determine the pile driving parameters of the pile foundation according to the preset marking points and the vertical pixel calibration value;

[0194] Input the pile driving parameters of the pile foundation into the bearing capacity prediction model to obtain the predicted bearing capacity of the pile foundation to be monitored.

[0195] In addition, when the logical instructions in the above-mentioned memory 330 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0196] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the monitoring method for pile driving of the pile foundation provided by the above-mentioned various methods. The method includes:

[0197] Obtain an image of pile driving of the pile foundation through a collection device set at a preset position. The image of pile driving of the pile foundation includes the pile foundation to be monitored, and a pile body scale is provided on the pile foundation to be monitored;

[0198] Calculate the actual size of a unit pixel according to the pile body scale to determine the vertical pixel calibration value;

[0199] Determine the pile driving parameters of the pile foundation according to the preset marking points and the vertical pixel calibration value;

[0200] Input the pile driving parameters of the pile foundation into the bearing capacity prediction model to obtain the predicted bearing capacity of the pile foundation to be monitored.

[0201] On yet another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the monitoring method for pile driving of the pile foundation provided by the above-mentioned various methods. The method includes:

[0202] Obtain the image of pile foundation sinking by means of a collection device set at a preset position. The image of pile foundation sinking includes the pile foundation to be monitored, and a pile body scale is provided on the pile foundation to be monitored.

[0203] Calculate the actual size of a unit pixel according to the pile body scale to determine the vertical pixel calibration value.

[0204] Determine the pile foundation sinking parameters according to the preset marking points and the vertical pixel calibration value.

[0205] Input the pile foundation sinking parameters into the bearing capacity prediction model to obtain the predicted bearing capacity of the pile foundation to be monitored.

[0206] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0207] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course also by hardware. Based on this understanding, the above technical solution, in essence, or the part 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, magnetic disk, optical disc, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0208] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or equivalently replace some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring pile foundation sinking, characterized in that: include: Acquire a pile foundation sinking image by means of a collection device arranged at a preset position, wherein the pile foundation sinking image includes a pile foundation to be monitored, and a pile body scale is arranged on the pile foundation to be monitored; Calculate the actual size of the unit pixel according to the pile body ruler to determine the vertical pixel calibration value; Determining pile foundation sinking parameters according to the preset marking points and the vertical pixel calibration values; Inputting the pile foundation sinking parameters into a bearing capacity prediction model to obtain the predicted bearing capacity of the pile foundation to be monitored; The preset marking points include tracking points, and the step of determining pile foundation sinking parameters according to the preset marking points and the vertical pixel calibration values ​​specifically includes: According to the current pile foundation sinking image, determine the vertical pixel coordinates of the current tracking point; According to the initial pile foundation sinking image, the vertical pixel coordinates of the initial tracking point are determined; Determine a pile penetration displacement variation curve according to the vertical pixel coordinates of each current tracking point and the vertical pixel coordinates of the initial tracking point, wherein the pile penetration displacement variation curve is used to characterize a variation curve of the pile penetration displacement over time; Determining the hammer penetration value and the hammer rebound value according to the pile penetration displacement variation curve; Wherein, the current tracking point and the initial tracking point are both set on the pile body scale; Determine the initial burial depth of the pile foundation to be monitored based on the scale value of the current tracking point, the absolute elevation of the scale position of the pile body where the tracking point is located, and the absolute elevation of the ground at the pile position; Determine the pile body penetration depth value during the construction process according to the initial penetration depth and the pile body penetration displacement change curve; Before the step of determining the pile foundation sinking parameters according to the preset marking points and the vertical pixel calibration values, the method further includes: When the tracking point coordinates exceed the preset threshold When , the tracking point position is updated according to the preset formula, and the reference sub-area matrix is ​​updated according to the tracking point position; The preset formula is as follows: in, is the vertical pixel coordinate of the updated tracking point, is the preset update distance, is the pixel calibration value, The vertical pixel coordinate of the current tracking point.

2. The method for monitoring pile foundation sinking according to claim 1, characterized in that: The step of calculating the actual size of a unit pixel according to the pile body scale to determine the vertical pixel calibration value specifically includes: Extracting a region including the pile body scale in the pile foundation sinking image; Extract the pixel positions of any two scale lines in the pile body scale in the image; The vertical pixel calibration value is determined according to the actual distance between the scale lines and the pixel distance, and the pixel distance is determined by the pixel position of each scale line in the image.

3. The method for monitoring pile foundation sinking according to claim 1, characterized in that: The pile foundation sinking parameters include hammer penetration, hammer rebound value and soil penetration depth.

4. The method for monitoring pile foundation sinking according to claim 1, characterized in that: After the step of acquiring the pile foundation sinking image by means of a collection device arranged at a preset position, the method further comprises: The pile foundation sinking image is preprocessed, and the preprocessing includes denoising, graying and contrast enhancement.

5. The method for monitoring pile foundation sinking according to claim 1, characterized in that: The bearing capacity prediction model is trained in the following way: Get historical pile foundation sinking parameters; Inputting the historical pile foundation sinking parameters into a pre-built bearing capacity prediction model for training until a convergence function is satisfied; The carrying capacity prediction model is constructed based on a multi-layer perception mechanism.

6. A monitoring device for pile foundation sinking, characterized in that: include: An image acquisition module, used to acquire a pile foundation sinking image through an acquisition device arranged at a preset position, wherein the pile foundation sinking image includes a pile foundation to be monitored, and a pile body scale is arranged on the pile foundation to be monitored; A calibration determination module, used for calculating the actual size of a unit pixel according to the pile body scale to determine a vertical pixel calibration value; A parameter determination module, used to determine pile foundation sinking parameters according to preset marking points and the vertical pixel calibration values; A bearing capacity prediction module, used for inputting the pile foundation sinking parameters into a bearing capacity prediction model to obtain the predicted bearing capacity of the pile foundation to be monitored; The preset marking points include tracking points, and the step of determining pile foundation sinking parameters according to the preset marking points and the vertical pixel calibration values ​​specifically includes: According to the current pile foundation sinking image, determine the vertical pixel coordinates of the current tracking point; According to the initial pile foundation sinking image, the vertical pixel coordinates of the initial tracking point are determined; Determine a pile penetration displacement variation curve according to the vertical pixel coordinates of each current tracking point and the vertical pixel coordinates of the initial tracking point, wherein the pile penetration displacement variation curve is used to characterize a variation curve of the pile penetration displacement over time; Determining the hammer penetration value and the hammer rebound value according to the pile penetration displacement variation curve; Wherein, the current tracking point and the initial tracking point are both set on the pile body scale; Determine the initial burial depth of the pile foundation to be monitored based on the scale value of the current tracking point, the absolute elevation of the scale position of the pile body where the tracking point is located, and the absolute elevation of the ground at the pile position; Determine the pile body penetration depth value during the construction process according to the initial penetration depth and the pile body penetration displacement change curve; Before the step of determining the pile foundation sinking parameters according to the preset marking points and the vertical pixel calibration values, the method further includes: When the tracking point coordinates exceed the preset threshold When , the tracking point position is updated according to the preset formula, and the reference sub-area matrix is ​​updated according to the tracking point position; The preset formula is as follows: in, is the vertical pixel coordinate of the updated tracking point, is the preset update distance, is the pixel calibration value, The vertical pixel coordinate of the current tracking point.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the pile foundation driving monitoring method as claimed in any one of claims 1 to 5 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for monitoring pile foundation sinking as claimed in any one of claims 1 to 5 is implemented.

Images