Intelligent Surveying and Mapping Method and Device for Pile Foundation Embedded Area Based on Fuzzy Coordinate Compensation

Abstract

The present invention provides an intelligent surveying and mapping method and device for the pre-buried area of pile foundations based on fuzzy coordinate compensation, which relates to the technical field of pile foundation surveying and mapping. The method includes steps of determining constraint conditions to generate a dynamic constraint association matrix; generating corresponding dense measurement points in the key node area of the pile foundation and the terrain mutation area respectively; outputting the error probability density characterizing the terrain error of the surveying and mapping area; deploying surveying and mapping equipment in the surveying and mapping area, preprocessing the collected data, and converting it into membership values through fuzzy processing; inputting the corrected collected data into the model, and re-estimating the model parameters through the EM algorithm to achieve high-precision construction optimization. By quantifying the strength values of the terrain mutation area and the key nodes of the pile foundation, the present invention dynamically adjusts the grid size and the measurement point density, so that the plane terrain and elevation errors are respectively controlled within ≤2 cm and ≤1 cm, the data redundancy rate is reduced by 30% - 50%, and the coverage rate of the key area is increased to more than 95%, accurately reflecting the terrain features.

Description

Technical Field

[0001] The present invention relates to the technical field of pile foundation surveying and mapping, and specifically to an intelligent surveying and mapping method and device for pile foundation embedded areas based on fuzzy coordinate compensation. Background Technique

[0002] The surveying and mapping of pile foundation embedded areas is a key technical link in modern civil engineering and infrastructure construction, and is widely used in the pre-construction planning and implementation stages of bridges, high-rise buildings, port facilities, and large-scale infrastructure. With the acceleration of the urbanization process and the advancement of large-scale infrastructure construction, the demand for surveying and mapping of pile foundation embedded areas is increasing day by day. At the same time, with the rapid development of unmanned aerial vehicle photogrammetry, point cloud data processing, and artificial intelligence technologies, intelligent surveying and mapping technologies have gradually become an important means to solve complex terrain surveying and mapping and high-precision construction requirements.

[0003] Although existing surveying and mapping technologies can provide a certain degree of accuracy, in complex terrain construction environments, especially in mountainous areas, wetlands, urban dense areas, and scenarios with high-precision construction requirements, they often face problems such as low efficiency, data redundancy, error accumulation, and insufficient adaptability to terrain mutations and pile foundation layouts. For example, in mountainous area surveying and mapping, due to terrain undulations and obstructions, the coverage of measurement points is uneven, resulting in insufficient data accuracy in key areas and data redundancy in non-key areas. Although unmanned aerial vehicle photogrammetry can quickly cover large areas, in areas with large vegetation obstructions and terrain undulations, the accuracy of its point cloud data is insufficient, and it is difficult to adapt to terrain mutations and dynamic changes in pile foundation layouts.

[0004] In addition, the measurement point planning method in traditional surveying and mapping relies on fixed grid division (or uniform grid division), and cannot dynamically adjust the measurement point density according to terrain complexity and engineering requirements. For example, in mountainous terrains, errors in flat areas are relatively concentrated, while errors in terrain mutation areas (such as steep slopes and valleys) are more dispersed. Using traditional methods will arrange too many measurement points in flat areas, resulting in data redundancy, while in terrain mutation areas, there are insufficient measurement points, and key terrain features and error distributions cannot be captured. This unreasonable distribution of measurement points not only increases the complexity of data processing, but also leads to insufficient measurement accuracy in key areas, affecting subsequent construction accuracy and engineering quality, resulting in measurement data being unable to accurately reflect terrain features, and further affecting the design and construction accuracy of bridge piers. Summary of the Invention

[0005] Aiming at the deficiencies of the existing technology, the purpose of the present invention is to provide an intelligent surveying and mapping method and device for pile foundation embedded areas based on fuzzy coordinate compensation, and solve the problems raised in the above background technique.

[0006] To achieve the above purpose, the present invention is realized through the following technical solutions: An intelligent surveying and mapping method for pile foundation embedded areas based on fuzzy coordinate compensation, including the steps:

[0007] Determine the constraint conditions and initialize the M×N matrix that matches the division dimension of the grid cell (i,j) in the survey area; use the Sigmoid function to calculate the terrain constraint intensity value T (i,j) and calculate the pile foundation constraint intensity distribution value P by combining the Gaussian kernel density (i,j) Generate a dynamic constraint correlation matrix through linear superposition and nonlinear correction , where α and β are weight coefficients used to measure the interactive influence of terrain and pile foundation constraints, and η is the interactive enhancement coefficient;

[0008] According to the normalized dynamic constraint correlation matrix, determine the measuring point density of each grid cell in the survey area. Among them, in the area with higher set constraint intensity, the measuring point density increases proportionally. Secondly, by comprehensively considering the slope and curvature data of all grid cells in the survey area and combining with the preset weight coefficient, calculate the terrain complexity estimation value of the current survey area, dynamically adjust the grid size required for surveying, and generate the measuring point layout logic that meets the said constraint conditions;

[0009] Deploy the device, generate dense coordinates based on the measuring point layout logic and collect elevation and point cloud data. After preprocessing, construct a structured data set, fuzzify it into low, medium, and high fuzzy sets through fuzzification processing, and use fuzzy clustering to extract the characteristics of the collected data;

[0010] Perform Gaussian component allocation on different terrains in the survey area, iterate and optimize through the EM algorithm until the model converges. After correcting the characteristics of the collected data as the input vector of the model according to the fuzzy rule base, input it into the model, and output the error probability density representing the terrain error in the survey area to achieve high-precision construction optimization.

[0011] As the second aspect of the present invention, an intelligent surveying device for pile foundation pre-buried area based on fuzzy coordinate compensation is proposed, including a memory and a processor. Among them, the memory includes an intelligent surveying program for pile foundation pre-buried area based on fuzzy coordinate compensation. When the intelligent surveying program for pile foundation pre-buried area based on fuzzy coordinate compensation is executed by the processor, it realizes the above-mentioned intelligent surveying method for pile foundation pre-buried area based on fuzzy coordinate compensation.

[0012] Compared with the prior art, the beneficial effects of the present invention are:

[0013] 1. To solve the problems in the prior art that the measurement point planning adopts static uniform distribution, resulting in insufficient data in terrain mutation areas, omission of key pile foundation nodes, and high data redundancy rate, the present invention constructs a constraint correlation matrix, quantifies the strength values of terrain mutation areas and key pile foundation nodes, dynamically adjusts the grid size and measurement point density, and encrypts measurement points in high-complexity areas, so that the plane terrain and elevation errors are respectively controlled within ≤2 cm and ≤1 cm, the data redundancy rate is reduced by 30% - 50%, and the coverage rate of key areas is increased to over 95%, accurately reflecting the terrain features;

[0014] 2. To solve the problem of insufficient accuracy in terrain error collection under complex terrain conditions, the present invention realizes data optimization through fuzzy inference correction. First, the key terrain features (standardized elevation value, point cloud density, elevation deviation) are fuzzified into membership values to generate a fuzzy vector, then fuzzy clustering is used to identify the key features of terrain data, the importance of feature clusters is evaluated and key features are extracted, then the correction weight is calculated through fuzzy inference, the elevation and point cloud density data are dynamically adjusted, and finally the corrected data is input into the Gaussian mixture model to update the model parameters, thereby improving the adaptability and prediction accuracy of the terrain error probability model and ensuring high-precision surveying and mapping and construction optimization under different terrain conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The disclosure of the present invention will be described with reference to the accompanying drawings. It should be understood that the drawings are only for illustrative purposes and are not intended to limit the scope of protection of the present invention. In the drawings, the same reference numerals are used to refer to the same components. Among them:

[0016] Figure 1 is a schematic flowchart of constructing an offline data collection strategy to ensure data collection accuracy proposed in an embodiment of the present invention;

[0017] Figure 2 is a schematic flowchart of constructing a terrain error probability model to control construction errors within the millimeter range proposed in an embodiment of the present invention;

[0018] Figure 3 is a schematic flowchart of using fuzzy clustering to optimize the input elevation and point cloud data, finally updating the terrain error probability model, and improving the surveying and mapping accuracy proposed in an embodiment of the present invention;

[0019] Figure 4 is a schematic flowchart of realizing high-precision construction optimization proposed in an embodiment of the present invention;

[0020] Figure 5 is a schematic diagram of the overall structure of a total station surveying instrument proposed in an embodiment of the present invention;

[0021] Figure 6This is a schematic structural diagram of a surveying triangular support frame and a rectangular displacement component proposed in an embodiment of the present invention;

[0022] Figure 7 This is a schematic exploded view of a rectangular displacement component proposed in an embodiment of the present invention.

[0023] Reference numerals:

[0024] 1. Surveying triangular support frame; 11. Central circular rod frame; 111. Lower sleeve ring; 112. Upper sleeve ring; 12. Support angle adjustment component; 121. Thick socket rod; 122. Angle engaging block; 123. Angle support rod; 13. Height adjustment component; 131. Thin socket rod; 1311. Support ball; 1312. Insertion pin; 132. Limit engaging block;

[0025] 2. Rectangular displacement component; 21. Displacement rectangular frame; 211. Lower linear groove; 212. Upper linear groove; 213. Rectangular rod; 2131. Limit opening; 22. Limit cover; 221. Centering mark; 222. Cross-shaped groove; 223. Side bayonet;

[0026] 3. Level; 4. Surveying instrument assembly. Detailed implementation manners

[0027] It is easy to understand that according to the technical solution of the present invention, without changing the essence of the present invention, those of ordinary skill in the art can propose various interchangeable structural ways and implementation manners. Therefore, the following detailed implementation manners and the accompanying drawings are only exemplary descriptions of the technical solution of the present invention, and should not be regarded as all of the present invention or as a limitation or restriction on the technical solution of the present invention.

[0028] The present invention will be further described in detail below with reference to the accompanying drawings, but it is not a limitation to the present invention.

[0029] As an understanding of the technical concept of the present invention, traditional measurement point planning adopts static uniform distribution, which is difficult to adapt to complex terrains (such as sudden slope changes and dense pile foundation areas), resulting in data redundancy (too dense in flat areas) or omission of key areas (insufficient in mutation areas). Therefore, the present invention proposes to generate a dynamic grid by quantifying the constraints of terrain mutation areas (slope - Sigmoid mapping to intensity values) and pile foundation nodes (Gaussian kernel density estimation). For example, in mountain surveying, the measurement point density in areas with a slope > 15° is increased to 2 times to ensure capturing terrain undulations, and the grid within a radius of 2d around the pile foundation is compressed to Δ final / 2, accurately covering the pile foundation coupling effect. Next, in order to capture the key terrain features and error distribution, the present invention allocates Gaussian components according to terrain types (flat areas, mutation areas, pile positions) to capture the error distribution of different areas, and dynamically adjusts weights, means, and covariances through expected maximization iterations to effectively deal with the skewed error distribution caused by construction disturbances in the pile position area, so as to achieve the effect of accurately quantifying the scope of construction influence. Finally, in order to make the measurement data accurately reflect the terrain features, the present invention converts continuous variables (elevation, point cloud density) into fuzzy sets (such as "high deviation" and "low density"), eliminates dimensional differences, and dynamically adjusts error model parameters based on expert knowledge or data-driven rules (such as "IF high deviation AND low density THEN increase compensation weight") to improve the accuracy and reliability of complex terrain mapping.

[0030] As an embodiment of the present invention, a method for intelligent mapping of a pile foundation pre-embedded area based on fuzzy coordinate compensation is first proposed, comprising the following steps:

[0031] like Figure 1 As shown in S1, an offline data acquisition strategy is constructed to complete the grid division and measurement point coordinate location planning of the terrain and pile foundation coupling in the surveying area (pile foundation pre-buried area) under different terrain conditions, so as to ensure that the terrain mutation area and the key nodes of the pile foundation layout are accurately covered during the subsequent surveying and mapping, and data redundancy is avoided. The implementation steps include:

[0032] S1-1. According to the design requirements of the infrastructure project, determine the terrain mutation area with a slope value greater than 15° and the pile foundation coordinates (x k ,y k ) The constraint conditions include the key node area of ​​the pile foundation group with a surrounding radius of n=2 times the average spacing d of the pile foundation. It can be understood that the purpose of determining the constraint conditions is to guide the reasonable layout of the measuring points and ensure that more measuring points can be allocated in areas with high terrain complexity (such as areas with large slopes or drastic changes in curvature), thereby improving the measurement accuracy. Through the constraint conditions, the grid size and measuring point density can be dynamically adjusted to generate a dense set of measuring point coordinates in key areas (such as key nodes of pile foundations and terrain mutation zones), ensuring that the acquisition accuracy and elevation error of the surveying and mapping data are controlled within the range of engineering design requirements, while optimizing resource allocation, avoiding excessive layout of measuring points in flat or simple terrain areas, and improving surveying and mapping efficiency.

[0033] Initialize an M×N matrix and make its dimension match the grid unit (i, j) division of the survey area; where the terrain mutation area calculates the slope value S of each grid unit through the DEM data model (i,j) Determined, if the slope value S(i,j) Greater than the slope threshold S th = 15°, it is regarded as a terrain mutation area. Assuming that the survey area is divided into M×N grid cells, the size of the constraint correlation matrix is also M×N; the radius of 2d around the pile foundation coordinates means that in pile foundation construction, a circular area is demarcated with the coordinates of each pile foundation position as the center and n = 2 times the average pile spacing d as the radius. This area is used to define the key nodes of the pile group, ensuring that during the surveying and construction processes, the terrain of the pile position and its surrounding areas can be measured and analyzed in detail, thereby guaranteeing the accuracy and stability of pile foundation construction.

[0034] Use the Sigmoid function to map the terrain mutation area (slope > 15°) to a continuous intensity value T (i,j) (0 to 1), and quantify the discrete terrain features into continuous data. Through this mapping, the influence degree of the slope on the construction difficulty or measurement accuracy can be more intuitively represented. , where S (i,j) is the actual slope measurement value at the grid cell (i,j) in the survey area, S th = 15°, is the slope threshold, r = 0.5, is the slope sensitivity parameter used to control the steepness of the Sigmoid function, T (i,j) is used to reflect the influence of the slope on the construction difficulty, and the larger the value, the more complex the terrain;

[0035] Based on the obtained pile foundation coordinates (x k , y k ) and the average pile spacing d, quantify the influence of the density of the key nodes of the pile group on the grid cell (i,j) through the Gaussian kernel density, and normalize to generate the pile foundation constraint strength distribution value P (i,j)： ， where, (x k , y k ) is the coordinate position of the kth pile foundation, k = 1, 2,..., K, and K is the total number of pile foundations. It can be understood that by calculating the density relationship between each grid cell and the pile position through the Gaussian kernel function, the purpose is to reflect the distribution of the pile foundation in the spatial area of the survey area. In order to make the results more comparable, it is also necessary to normalize the calculated values so that their values are between 0 and 1: , where max (p,q) is the maximum value in the entire pile foundation constraint strength distribution, used for normalization processing to ensure that all values are scaled to the range of 0 to 1, P (p,q) is the original value at the position of a certain grid cell (i,j), and p, q are index variables. Through this quantification process, the constraint conditions of the key nodes of the pile group can be transformed into continuous intensity values, providing data support for subsequent measurement point planning.

[0036] The continuous strength value T (i,j) and the pile foundation constraint strength distribution value P (i,j) are linearly superimposed, and a non-linear correction factor is introduced to construct a dynamic and unnormalized constraint correlation matrix to achieve the surveying and mapping logic that needs to be focused on in complex terrain - dense pile groups in infrastructure projects: , where α and β are weight coefficients used to measure the interactive influence between terrain and pile foundation constraints, usually taking values α = β = 1, and η is an interactive enhancement coefficient, usually taking a value of 0.5. After completing the matrix construction, the preliminary distributed measurement point density ρ(i,j) at the position of each grid cell (i,j) is output to understand the distribution of measurement point coordinates under different terrain conditions, ensure that the measurement points can accurately cover the terrain mutation area and the key nodes of the pile foundation, and at the same time avoid data redundancy in non-critical areas, improve the efficiency and accuracy of subsequent surveying and mapping work, and provide reliable measurement point data for the subsequent construction of the terrain error probability model and pile foundation pre-embedding construction. The calculation steps of the measurement point density ρ(i,j) are as follows:

[0037] First, linearly map the constraint correlation matrix W raw to the range from 0 to 1: , where max (raw) and min (raw) are the maximum and minimum values in the constraint correlation matrix respectively, and W nor (i,j) is the value of the grid cell (i,j) in the normalized constraint correlation matrix;

[0038] Secondly, for areas that simultaneously meet high terrain strength (such as T (i,j) > 0.8) and high pile foundation strength (such as P (i,j) > 0.7), increase their weight values to achieve the priority adjustment of conflict areas: W nor (i,j) = min(1.0, W nor (i,j) + 0.2);

[0039] Finally, calculate the measurement point density of each grid cell (i,j) according to the normalized constraint correlation matrix: ρ(i,j) = ρ base ×(1 + h×W nor (i,j)), where ρ base is the basic measurement point density, ρ base = 1 / Δ 2, represents the expected number of basic measuring points within the unit grid area, the coefficient h=3, is the amplification factor of the constraint strength on the measuring point density, which is used to enhance the influence of the constraint strength on the measuring point density, Δ is the basic grid size divided according to the terrain complexity estimate C and the average spacing d of the pile foundation, Δ=d / (k×C), k is the adjustment coefficient, which is determined according to the terrain complexity and the pile foundation density. In this embodiment, in areas with complex terrain or dense pile foundations, k=1.5 is taken to reduce the grid size and increase the measuring point density. In areas with simple terrain or sparse pile foundations, k=2.0 is taken to increase the grid size and improve the collection efficiency.

[0040] S1-2. Calculate the terrain complexity estimate C according to the terrain characteristics of the survey area to quantify the complexity of the terrain. Then, dynamically adjust the grid size Δ based on the terrain complexity estimate C. final It should be noted that the terrain complexity estimation C needs to comprehensively consider the two factors of slope S and curvature K for comprehensive calculation: , where is the slope value of the i-th measuring point, reflecting the inclination of the terrain in the surveying area. The larger the absolute value, the steeper the slope. is the curvature value of the i-th measuring point, reflecting the curvature of the terrain in the surveying area. The larger the absolute value, the more drastic the terrain change. The maximum slope value in the survey area is used to normalize the slope so that the slopes at different measuring points are comparable. is the maximum curvature value in the surveying area. a and b are the weight coefficients of slope and curvature respectively. According to experience, the slope weight is 40% and the curvature weight is 40%. The remaining 20% ​​is adjusted according to other factors in actual operation (such as elevation change rate).

[0041] Dynamically adjust the grid size Δ final The formula is: , where Δ is the basic grid size divided according to the terrain complexity estimate C and the average spacing d of pile foundations, ρ base is the basic measurement point density, ρ(i,j) is the measurement point density, η is an adjustment index used to control the influence of the measurement point density on the grid size, usually between 0.5 and 2, and the specific value is determined according to the actual terrain complexity and engineering requirements, ξ is an adjustment coefficient used to further fine-tune the grid size, which is set according to the actual measurement accuracy requirements and resource constraints, and usually between 0 and 1. So far, by dynamically adjusting the grid size Δ final , reduce the grid size in areas with complex terrain and high density, and increase the size conversely, and simultaneously optimize the distribution of measurement point density to ensure that the measurement point density in high-complexity areas is increased to 2 times, and the density in low-complexity areas is reduced to avoid data redundancy, and finally generate an adaptive grid division scheme to balance the subsequent surveying and mapping accuracy and efficiency.

[0042] S1-3. According to the adjusted grid size Δ final , divide the survey area into multiple grid cells, and generate a measuring point layout logic that meets the constraint conditions. Among them, set the size of each grid cell to be Δ final ×Δ final . Evenly distribute measuring points within each grid cell according to the measuring point density ρ(i,j) to capture the topographic features around the pile foundation in detail. The specific measuring point layout logic is as follows:

[0043] Within the key node area of the pile group with a radius of 2 times the average pile spacing d around the pile foundation coordinates (x k , y k ), double the grid density, with the grid size being Δ final / 2. Generate the coordinates of dense measuring points according to the following formula: , where k is the index of the measuring point, ranging from 0 to ρ(i,j)-1. That is, the dense measuring points are alternately offset based on the starting point according to the parity of the index k, with a horizontal interval of Δ final / 2 and increasing by a fixed step size in the vertical direction. x start is the starting coordinate of the pile foundation coordinate in the x-axis direction, and y start is the starting coordinate of the pile foundation coordinate in the y-axis direction. Usually, it represents the starting position or reference point of the pile foundation coordinate, which is the basis for generating the coordinates of dense measuring points and is used to determine the starting positions of dense measuring points in the horizontal and vertical directions.

[0044] In the terrain mutation area where the slope value is greater than 15°, adjust the positions of the measuring points according to the local curvature K to increase the measuring point density. The greater the curvature, the greater the offset amplitude of the measuring points, and the offset direction is determined by the components of the curvature in the horizontal and vertical directions, so as to ensure that the measuring point density increases with the improvement of terrain complexity and avoid data distortion caused by excessive offset. Generate the coordinates of dense measuring points according to the following formula: , where is the adjusted coordinate of the measuring point, , is a fixed basic offset, usually taken as 0.5 meters, to ensure that the adjustment amplitude is within a reasonable range, , are the components of the local curvature in the x and y directions and are used to adjust the positions of the measuring points according to the terrain curvature.

[0045] S1-4. Finally, output the set of coordinates of dense measuring points to ensure that the data acquisition accuracy of the survey area reaches a plane error ≤ 2 cm and an elevation error ≤ 1 cm, meeting the engineering design requirements.

[0046] Such as Figure 2As shown in the figure, as an embodiment of the present invention, the proposed intelligent surveying and mapping method further includes the following steps:

[0047] S2. Construct a terrain error probability model.

[0048] It should be noted that the terrain error probability model is constructed based on the Gaussian mixture model (GMM). The purpose is to accurately quantify the statistical characteristics and spatial correlations of terrain errors in terrain surveying and pile foundation construction. By fusing elevation data and point cloud data, the model can capture the subtle changes and complex features of the terrain. Especially in the pile foundation embedding area, these areas usually have a high point cloud density and significant elevation deviations. This model generates the error probability density and risk identification matrix through initializing Gaussian component parameters, iterative optimization, and verification output. The output data can directly support the real-time decision-making and automatic compensation of the construction management system, ensuring that the construction accuracy meets the engineering requirements. For example, in the pile foundation embedding area, the model can identify high-probability error areas, trigger construction alarms, and generate compensation strategies, thereby controlling the construction error within the millimeter range and significantly improving construction efficiency and safety.

[0049] It can be understood that in the embodiment of the present invention, terrain error specifically refers to the measurement error caused by terrain features (such as slope, curvature, terrain mutation, etc.), which is mainly related to terrain changes and is the difference between the measured value of the device and the true terrain value caused by various factors. These errors come from equipment accuracy limitations, terrain complexity, environmental interference, or data processing methods, etc. As shown in Table 1, it is several common terrain error types and their specific manifestations in infrastructure projects:

[0050] Table 1

[0051]

[0052] The specific implementation steps of S2 include:

[0053] S2-1. Initialization of the Gaussian mixture model (GMM):

[0054] In the model initialization stage, first, Gaussian components are assigned to different terrains in the surveyed area. For the pile position area, M1 = 3 Gaussian components are assigned, and their means are initialized to the centers of the encrypted grids, and their covariance matrices are set as diagonal matrices, including (Δ / 4) 2 , (Δ / 4) 2 and σ e 2; For the mutation area, assign M2 = 2 components, whose means are weighted by curvature; for the flat area, assign M3 = 1 component, and set its covariance matrix to the maximum value to simplify the model. It can be understood that to quantify the multimodal distribution and spatial correlation of terrain errors, Gaussian components need to be assigned according to terrain types (flat area, mutation area, pile position area) to capture the terrain data errors in different survey areas. In practice, the high-density measurement point data in the pile position area (step S1) enables the GMM to assign 3 Gaussian components, which can respectively capture the pile foundation construction disturbance, soil bearing capacity difference, and peripheral micro-topography coupling effect.

[0055] Secondly, perform the EM algorithm iteration. It should be noted that in the iterative optimization process of the EM algorithm, first execute the E-step (expectation step) to calculate the posterior probability that each grid cell belongs to each Gaussian component. The posterior probability is realized by substituting the current weights, means, and covariance matrices into the Gaussian distribution density function, which is used to measure the matching degree between the input vector and each Gaussian component, and the posterior probability is used to reflect the possibility that each grid cell data point belongs to each Gaussian component. It can be understood that the covariance matrix output by the GMM quantifies the spatial correlation of errors (such as the error similarity of adjacent grids), providing a physical basis for the subsequent fuzzy correction rules. If the GMM shows a high spatial correlation of errors in a certain area, the fuzzy rule will preferentially perform overall compensation rather than local adjustment on this area.

[0056] Thirdly, perform the M-step (maximization step), and update the model parameters in turn to maximize the log-likelihood function: first update the weights, and obtain the new weight value ωj by summing and normalizing the posterior probabilities belonging to each Gaussian component; then update the means, by multiplying the input vector of each grid cell by the posterior probability and summing, and then dividing by the total sum of the posterior probabilities of this Gaussian component; finally update the covariance matrix, by calculating the outer product of the difference between the input vector and the new mean, multiplying by the posterior probability and summing, and then normalizing.

[0057] Finally, set the model convergence condition: that is, stop the iteration when the change in the log-likelihood value between two consecutive iterations is less than the system threshold (10 to the power of negative 5) to ensure that the model parameters are stable and optimal, so as to optimize the parameters of the Gaussian mixture model, make it better fit the data, and accurately quantify the probability distribution of terrain errors.

[0058] S2-2, verify the model and generate the error probability density: perform log-likelihood evaluation to select the optimal number of Gaussian components; conduct residual analysis to verify the rationality of the model; output structured error probability density data for quantifying terrain errors, including Gaussian component parameters, error probability density, and risk identification matrix. It can be understood that the risk identification matrix (such as marking the high-risk area as 1) directly guides the construction path planning in step S4.

[0059] It should be noted that during the model validation and application stage, when performing model validation to ensure its effectiveness, for log-likelihood evaluation, the log-likelihood value of the test set is calculated to evaluate the model fitting effect, and then (such as through the BIC criterion) the number of Gaussian components M that maximizes the log-likelihood value is selected as the optimal number of Gaussian components; for residual analysis, the residual between the predicted elevation and the actual elevation is calculated to verify whether it follows a normal distribution with a mean of 0. Specifically, during implementation, it is verified through a Q-Q plot or a K-S test. The commonly used software tool preferably selects the scikit-learn library of Python for the implementation of the Gaussian mixture model and the calculation of the BIC criterion, and preferably uses the SciPy library of Python for Q-Q plot drawing and K-S test. For the error probability density, the error probability density of each grid cell is calculated through the model to form an error probability density map. Specifically, during implementation, this step preferably uses the Matplotlib or Seaborn library of Python for visualization to intuitively display the error distribution. Based on the NumPy library of Python, high-risk areas (error probability greater than 0.9) are marked as 1, and other areas are marked as 0 to quickly generate a risk identification matrix.

[0060] In an embodiment of the present invention, the error probability density calculation formula is: , where M is the number of Gaussian components in the Gaussian mixture model (GMM). The model consists of M Gaussian components, and the weight of each component is ωj, the mean is , and the variance is . It can be understood that the contribution of each Gaussian component is jointly determined by its weight and the Gaussian distribution density function. The weight reflects the importance of the component in the overall model, while the Gaussian distribution density function describes the distribution of errors under this component. In this embodiment, ωj is the weight of the jth Gaussian component, indicating the importance of the component in the overall model, is the Gaussian distribution density function, indicating the probability density of the error e under the jth Gaussian component, is the mean of the jth Gaussian component, indicating the central position of the component, is the variance of the jth Gaussian component, indicating the distribution range of the component.

[0061] As Figure 3 shown, as an embodiment of the present invention, the proposed intelligent mapping method further includes the following steps:

[0062] S3. Update the terrain error probability model to improve the mapping accuracy (plane error ≤ 2 cm, elevation error ≤ 1 cm). The implementation steps include:

[0063] S3-1. Deploy surveying equipment in the surveying area (pre-buried pile foundation area) and obtain original terrain data that accurately covers the terrain mutation area and the key node area of ​​the pile foundation through the coordinates of the measuring points. The original terrain data includes: elevation data Z that characterizes the basic terrain attributes of the surveying area collected by the total station surveying instrument, GPS equipment and drone photogrammetry equipment i,j And point cloud data PC i,j In this embodiment, high-density measurement is performed in the area surrounding the pile foundation (such as radius r=2d) to obtain elevation data Z i,j , with an accuracy of ±2 mm, the GPS device used is set to provide global geographic positioning, ensuring that the coordinate system is WGS84, and the coverage error is controlled within ±5 cm. When the drone photogrammetry equipment is used for measurement: point cloud data is collected through a multispectral sensor PC i,j , with a resolution of 0.1 m, focusing on covering terrain mutation areas with slope S>15∘.

[0064] S3-2, outlier removal: remove outliers caused by noise from surveying equipment or environmental interference (such as vegetation occlusion), use the 3σ principle to filter outliers in elevation data, and clean invalid points in point cloud data through statistical filtering. In this embodiment, when removing outliers, first perform an analysis on the elevation data Z. i,j Filter and calculate the global elevation mean μ Z and standard deviation σ Z , and then remove those exceeding μ Z ±3σ Z The elevation values ​​within the range of three standard deviations around the mean are retained. In practice, elevation values ​​outside this range are considered outliers and are removed. , , where M and N are the number of rows and columns of the grid, respectively, which determines the total number of grids to be MN.

[0065] For point cloud data cleaning, radius filtering technology (such as Radius Outlier Removal) is used to calculate the distance dk from each point cloud data to the center of the grid where it is located. If the distance dk exceeds twice the final grid size 2Δ final , the point cloud data point is considered to be an isolated point and is removed from the data set. In this embodiment, isolated points with less than 5 points in the neighborhood are removed. This ensures that the retained point cloud data is reasonable in spatial distribution and removes abnormal points caused by measurement errors or environmental interference. It can be understood that through the above steps, the outliers and invalid points in the original terrain data can be effectively cleaned up, thereby improving the accuracy and reliability of the terrain error probability model finally constructed.

[0066] S3-3. Dynamically adjust the grid size to adapt to terrain complexity and pile foundation layout. For example, compress the grid in the pile position area to increase data density and construct a structured data set D. The specific implementation steps are as follows:

[0067] S3-31. Dynamically adjust the grid size according to terrain complexity and pile foundation layout: For the pile position area, compress the grid size to Δ final / 2 to double the measurement point density; for the terrain mutation area (slope S>15°), keep the grid size Δ final = 1 m; for the flat area (slope S<15°), expand the grid size to Δ final = 2 m;

[0068] S3-32. Next, perform coordinate mapping to ensure that the measurement points are accurately mapped into the corresponding grid cells. Map the measurement point coordinates to the grid index (i,j): , , where and are the minimum coordinates of the survey area.

[0069] S3-33. Perform Z-score standardization on the elevation data to construct a structured data set D, D={( Z' i,j ,PC xc , yc .las)}, which contains the standardized elevation data Z' i,j and the corresponding point cloud file PC xc , yc .las. Among them, the point cloud file is named according to the coordinates of the grid center, such as PC105.535.5.las, so as to associate its position information.

[0070] S3-4. In the structured data set D, extract the standardized elevation value Z' i,j , point cloud density N i,j (the number of point clouds in the grid), and elevation deviation e i,j (the absolute value of the difference between the measured value and the theoretical value). Construct the model input vector V i,j , V i,j = [Z' i,j , N i,j , e i,j for subsequent Gaussian mixture model (GMM) training and optimization to quantify the spatial distribution of terrain errors. Specifically, for the point cloud density N i,j , based on the above coordinate mapping steps, after mapping the point cloud data into the grid cells, count the number of point clouds in each grid cell (i,j), which is denoted as the point cloud density N i,j ; for the elevation deviation ei,j , calculated by averaging the deviation between the elevation of the point cloud within the quantization grid cell (i, j) and the standardized elevation value: , where Zp is the elevation value of the p-th point cloud point, representing the actually measured elevation data.

[0071] S3-5. Conduct feature selection and fuzzification processing, and transform the model input vector V i,j , V i,j = [Z' i,j , N i,j , e i,j into a fuzzy set, convert it into a quantifiable membership value, and generate a fuzzy vector F i , F i = [μ 低 (Z' i,j ), μ 中 (N i,j ), μ 高 (e i,j ), …], thereby quantifying the influence degree of each feature data on the terrain error. Among them, for the standardized elevation value Z' i,j , construct a triangular membership function, and fuzzify it into three fuzzy sets of "low elevation", "medium elevation", and "high elevation", corresponding to different elevation ranges respectively, so as to reflect the influence of elevation data on the terrain error. In this embodiment, for the low elevation membership function in the standardized elevation value Z' i,j is defined as follows: when Z' i,j is in the interval of [Z min , Z mid1 , the membership degree decreases linearly with the increase of Z', until it drops to 0 at Z i,j , and when it exceeds this interval, the membership degree is always 0. For the medium elevation membership function in the standardized elevation value Z' mid1 is defined as follows: in the interval of [Z i,j , Z min , Z mid1 , the membership degree increases linearly with the increase of Z', and in the interval of (Z i,j , Z mid1 , Z mid2 , the membership degree decreases linearly with the increase of Z', and when it exceeds this range, the membership degree is 0. For the high elevation membership function in the standardized elevation value Z' i,j is defined as follows: when Z' i,j is in the interval of [Z i,j , Z mid2 , Z max , the membership degree increases linearly with the increase of Z', until it reaches 1 at Z i,j , and when it exceeds this interval, the membership degree remains 1 (if Z' max ​i,j Greater than Z max ) or always 0 (if Z' i,j Less than Z mid2 ).

[0072] For the point cloud density N i,j , it is fuzzified into "low density", "medium density", and "high density" through a trapezoidal membership function to quantify the influence of the reliability of point cloud density data on the terrain error. In this embodiment, when the point cloud density N i,j is less than or equal to the low density threshold N low , the membership degree is 1. When N i,j is in the interval (N low , N mid , the membership degree decreases linearly with the increase of density until it drops to 0 at N mid . When it exceeds this interval, the membership degree is always 0. In the interval [N low , N mid , the membership degree increases linearly with the increase of density. In the interval (N mid , N high , the membership degree decreases linearly with the increase of density. When it exceeds this range, the membership degree is 0. When the density N i,j is in the interval [N mid , N high , the membership degree increases linearly with the increase of density until it reaches 1 at N high . When the density exceeds N high , the membership degree remains 1. If the density is lower than N mid , the membership degree is always 0.

[0073] For the elevation deviation e i,j , it is fuzzified into "low deviation", "medium deviation", and "high deviation" using a triangular membership function to directly reflect the degree of measurement error. In this embodiment, for the elevation deviation e i,j , the low deviation membership function is defined as follows: When the elevation deviation e i,j is in the interval [0, e low , the membership degree decreases linearly with the increase of deviation until it drops to 0 at e low . When it exceeds this interval, the membership degree is always 0. The medium deviation membership function: In the interval [0, e mid , the membership degree increases linearly with the increase of deviation. In the interval (e mid , e high , the membership degree decreases linearly with the increase of deviation. When it exceeds this range, the membership degree is 0. The high deviation membership function: When the deviation e i,j is in the interval [e high , e max , the membership degree increases linearly with the increase of deviation until it reaches 1 at e maxReaches 1 at a certain point, and when the deviation exceeds e max the membership degree remains 1. If the deviation is lower than e high the membership degree is constantly 0.

[0074] Based on the above technical concept, it can be understood that by fuzzifying the above features, the model can handle complex terrain data more flexibly, achieve refined processing of terrain errors, and thus improve the adaptability and prediction accuracy of the model under complex terrain conditions.

[0075] S3-6. Construct a membership matrix to identify the key features of terrain data. It can be understood that the key features of terrain data include: flat areas: the terrain is relatively flat, with a small slope and low curvature. In actual infrastructure construction, the characteristics of these areas are gentle elevation changes and uniform data distribution, and usually do not require a high density of measurement points; terrain mutation areas: the terrain changes violently, with a large slope (such as a slope greater than 15°) and high curvature. In actual infrastructure construction, these areas need special attention because they have a significant impact on construction accuracy and stability, and usually require an increase in the measurement point density to capture terrain details; pile position areas: the key areas for pile foundation construction, usually the area with a radius of 2d centered on the pile foundation coordinates. In actual infrastructure construction, these areas require a high density of measurement points to ensure the accuracy of pile foundation construction. The specific implementation steps include:

[0076] S3-61. Construct the objective function to minimize the clustering algorithm: , where N is the total number of grids, representing the number of grids divided in the entire survey area, C = 3 is the number of clusters, representing flat areas, mutation areas, and pile position areas respectively. It should be noted that the three terrain types have different correction requirements in the intelligent survey of the actual pile foundation embedding area. u ic is the membership degree of grid i to cluster c, representing the degree to which grid i belongs to cluster c, satisfying ∑ c=1 C u ic = 1. The purpose is to identify the distribution of different terrain features in the pile foundation embedding area. m = 2 is the fuzzy index, which is used to describe the fuzziness of features such as normalized elevation values, point cloud density, and elevation deviation, so as to control the fuzziness of the membership degree. It can be understood that a larger fuzzy index allows a grid to belong to multiple clusters simultaneously, which is particularly important for dealing with the transition zones of terrain features in the pile foundation embedding area. vc is the center vector of cluster c, representing the core features of the cluster, providing a basis for subsequent evaluation of the importance of feature clusters and construction of the fuzzy rule base. ||·|| is the Euclidean distance, which is used to measure the distance between the fuzzy vector Fi and the cluster center vc to ensure the accuracy and reliability of the clustering results.

[0077] S3-62. Construct the membership update formula to update the cluster center and identify the key features of the collected terrain data. Specifically, when implementing, the membership update formula is: , where v d is the central vector of cluster d.

[0078] Cluster center update: ; To avoid infinite iteration, in implementation, the following comprehensive termination conditions are preferentially set to ensure the stability and efficiency of the clustering process: Set a maximum number of iterations. For example, when the number of iterations reaches Tmax = 100, the iteration stops regardless of whether the clustering is completely convergent. Or stop when any one of the conditions of change in the membership matrix, change in the cluster center, or change in the objective function is satisfied. Among them, the condition for change in the membership matrix: The change in the membership matrix U in two consecutive iterations is less than the system-set threshold; the condition for change in the cluster center: The change in the cluster center vc in two consecutive iterations is less than the system-set threshold; the condition for change in the objective function: The change in the objective function J in two consecutive iterations is less than the system-set threshold.

[0079] The membership matrix U is obtained, U = [u ic N×C， to represent the membership of each grid to each cluster; the cluster center vectors {v1, v2, v3} respectively correspond to the core features of the flat area, mutation area, and pile position area.

[0080] Based on the above technical concept, it can be understood that by minimizing the objective function J and continuously optimizing the membership matrix U and the cluster center vc in the clustering process, and finally outputting the membership matrix U and the cluster center vc, a refined processing of terrain errors can be achieved in the subsequent intelligent mapping of the pre-buried area of the pile foundation.

[0081] S3-7. It should be noted that in terrain mapping, due to the complexity and diversity of terrain features, such as flat areas, terrain mutation areas, and pile position areas, traditional measurement methods and models often have difficulty accurately capturing terrain changes, resulting in errors in measurement data. These errors will affect construction accuracy and project costs. Therefore, in this embodiment, a method for evaluating the importance of feature clusters is further proposed to quantify the influence of each feature on terrain errors, ensure that the measurement points can accurately cover the terrain mutation area and the key nodes of the pile foundation, and at the same time avoid data redundancy in non-critical areas. The specific implementation steps are as follows:

[0082] ​First, the compactness of feature clusters is evaluated by calculating the within-cluster variance. The smaller the variance, the higher the consistency of the data within the cluster, and the higher the importance score of the feature cluster. Second, the distinguishability between clusters is evaluated by calculating the distance between cluster centers. The larger the distance, the higher the distinguishability between clusters, and the higher the importance score of the feature cluster. Third, the correlation between each feature and the cluster center is calculated. Features with a correlation coefficient greater than 0.7 are selected as key features. It can be understood that these key features usually include the standardized elevation value, point cloud density, and elevation deviation, which have the greatest impact on terrain error. Finally, these key features are input into the fuzzy rule base constructed based on expert experience to provide a basis for subsequent terrain error correction, so as to ensure that the model can dynamically adjust the correction weight according to the changes in the actual measurement environment, improving the adaptability and prediction accuracy of the model.

[0083] In this embodiment, the specific steps to implement the correction of the collected data features are as follows:

[0084] First, the relationship between feature combinations and correction weights is defined through a fuzzy rule base constructed based on expert experience. For example, the rule "IF high deviation AND low density THEN high weight", where the triggering intensity of the rule is determined by the membership degrees of relevant features, such as the high membership degree of elevation deviation and the low membership degree of point cloud density. Second, a Mamdani fuzzy controller is used to infer and construct the output fuzzy set, and the specific correction weight is obtained by defuzzification using the centroid method. Finally, the correction weight is used to adjust the elevation and point cloud density to ensure that the output of the terrain error probability model is more accurate and adaptable to different terrain conditions.

[0085] S3-8. Update the terrain error probability model. The specific implementation steps are as follows:

[0086] The corrected elevation value and point cloud density are input into the Gaussian mixture model (GMM), and the model parameters, including weights, means, and covariance matrices, are re-estimated through the EM algorithm. The weights reflect the mixing ratio of the corrected error distribution, and the means and covariances describe the centers and distribution ranges of each Gaussian component. Using the updated parameters, the model can more accurately describe the probability distribution of terrain error, thereby improving the prediction accuracy.

[0087] As Figure 4 shown, as an embodiment of the present invention, the proposed intelligent surveying and mapping method further includes the following steps:

[0088] S4. Based on the updated terrain error probability model, achieve high-precision construction optimization. The specific implementation steps are as follows:

[0089] Use AutoCAD Civil 3D to generate a sub-meter terrain surface, extract the pile foundation boundary in combination with the OpenCV library, and construct a high-precision BIM model;

[0090] Through the NSGA-II algorithm and ROS mechanical path planning, with the goal of minimizing construction errors (relying on the terrain error probability model) and maximizing efficiency, dynamically avoid high-risk areas and recommend optimal pre-buried points;

[0091] With the help of the Unity digital twin platform, visually display construction data in real time, continuously update the error model and adjust the pre-buried points;

[0092] At the acceptance stage of the infrastructure project, compare the design model with the corrected terrain to generate a deviation report, ensuring that the plane error ≤ 2 cm and the elevation error ≤ 1 cm, and achieving accurate positioning of the pile foundation pre-buried area.

[0093] As the second aspect of the present invention, an intelligent surveying and mapping device for pile foundation pre-buried area based on fuzzy coordinate compensation is proposed, including a memory and a processor. Among them, the memory includes an intelligent surveying and mapping program for pile foundation pre-buried area based on fuzzy coordinate compensation. When the intelligent surveying and mapping program for pile foundation pre-buried area based on fuzzy coordinate compensation is executed by the processor, the intelligent surveying and mapping method for pile foundation pre-buried area based on fuzzy coordinate compensation proposed above is realized.

[0094] As Figures 5 - 7 shown, as an embodiment of the present invention, the total station surveying instrument proposed includes:

[0095] A surveying triangular support frame 1. A rectangular displacement component 2 for providing a moving guide rail is fixedly arranged at the upper end of the surveying triangular support frame 1. A spirit level 3 for adjusting the horizontal state is placed through the upper end of the rectangular displacement component 2. A surveying instrument component 4 for planning surveying is rotatably sleeved at the upper end of the spirit level 3.

[0096] In an embodiment of the present invention, the surveying triangular support frame 1 includes a central circular rod frame 11, a support angle adjustment component 12, and a height adjustment component. It can be understood that the central circular rod frame 11 is the core component of the surveying triangular support frame 1, playing a role in supporting and connecting the overall structure. An upper sleeve ring 112 is sleeved at its upper end, and a lower sleeve ring 111 is sleeved at its lower end. The distance between the upper sleeve ring and the lower sleeve ring can be adjusted by adjusting the height of the upper sleeve ring to adapt to different measurement requirements. The design purpose of the central circular rod frame 11 is to ensure the stability and flexibility of the support angle adjustment component 12 and the height adjustment component 13.

[0097] The support angle adjustment component 12 is a key part of the surveying triangle support frame, used to adjust the support angle to adapt to uneven ground conditions. There are multiple groups of support angle adjustment components 12, and each group of support angle adjustment components 12 includes a thick socket rod 121, an angle engaging block 122, and an angle support rod 123. Among them, the thick socket rod 121 is connected to the upper collar 112, the angle support rod 123 is connected to the lower collar 111, and the angle engaging block 122 is installed between the thick socket rod 121 and the angle support rod 123, used to fix and adjust the angle between the two. By adjusting the angle engaging block 122, the angle between the thick socket rod 121 and the angle support rod 123 can be changed, so as to adjust the support height. This design enables the theodolite to remain stable on the inclined ground and ensures the accuracy of measurement data from the perspective of physical parameters.

[0098] The height adjustment component 13 is used to finely adjust the overall height of the theodolite to adapt to different measurement environments. There are multiple groups of height adjustment components 13, and each group of height adjustment components includes a thin socket rod 131, a limit engaging block 132, a support ball 1311, and a pin 1312. Among them, the thin socket rod 131 is movably sleeved in the thick socket rod 121, and the height can be adjusted by telescoping and fixed with the limit engaging block 132. The support ball 1311 provides a stable support point, and the pin 1312 is inserted into the ground to ensure the stability of the structure, so that the theodolite can quickly adapt to different terrain conditions.

[0099] In an embodiment of the present invention, the rectangular displacement component 2 includes a displacement rectangular frame 21 and a limit cover 22. It can be understood that the displacement rectangular frame 21 is the core part of the rectangular displacement component, used to provide the movement of the level. There are mutually perpendicular lower linear grooves 211 and upper linear grooves 212 opened inside the displacement rectangular frame 21. The depths of the lower linear groove and the upper linear groove are different, ensuring that the level can move slightly within the horizontal range. There are rectangular rods 213 on both sides, and a limit opening 2131 is opened in the middle position, used to limit the level, ensuring that the level remains horizontal during the movement process and reducing measurement errors.

[0100] The limit cover 22 is fixedly arranged at the upper end of the displacement rectangular frame 21, used to limit the movement range of the level. Among them, a centered mark 221 is depicted or coated on the upper end of the limit cover 22 to mark the center position of the disc. A cross-shaped groove 222 is opened at the center of the disc, and side bayonets 223 are opened on both sides, used to limit the level, so as to ensure the precise calibration and stable movement of the level.

[0101] In an embodiment of the present invention, to ensure the precise movement and limit of the level, and prevent deviation during movement, a groove slider adapted to the displacement rectangular frame 21 and the cross-shaped groove 222 can be installed at its bottom, with side clamping blocks arranged around, so that the groove slider is movably clamped in the cross-shaped groove of the limit cover, and the side clamping blocks are slidably clamped in the side clamping openings 223.

[0102] The technical scope of the present invention is not limited to the content described above. Those skilled in the art can make various deformations and modifications to the above embodiments without departing from the technical idea of the present invention, and these deformations and modifications should all fall within the protection scope of the present invention.

Claims

1. An intelligent surveying and mapping method for the pre-embedded area of pile foundations based on fuzzy coordinate compensation, characterized in that: comprising the steps of: Determine the constraint conditions and initialize an M×N matrix that matches the division dimension of the grid cell (i,j) in the survey area; calculate the terrain constraint intensity value T using the Sigmoid function (i,j) and calculate the pile foundation constraint intensity distribution value P by combining the Gaussian kernel density (i,j) , generate a dynamic constraint correlation matrix through linear superposition and nonlinear correction , where α and β are weight coefficients used to measure the interactive influence of terrain and pile foundation constraints, and η is the interactive enhancement coefficient; According to the normalized dynamic constraint correlation matrix, determine the measuring point density of each grid cell in the survey area. Among them, in the area with higher set constraint intensity, the measuring point density is increased proportionally. Secondly, by comprehensively considering the slope and curvature data of all grid cells in the survey area and combining with the preset weight coefficient, calculate the terrain complexity estimation value of the current survey area, dynamically adjust the grid size required for surveying, and generate the measuring point layout logic that meets the said constraint conditions; Deploy devices, generate dense coordinates based on the measuring point layout logic, collect elevation and point cloud data, construct a structured data set after preprocessing, fuzzify it into low, medium, and high fuzzy sets through fuzzification processing, and use fuzzy clustering to extract the characteristics of the collected data; Perform Gaussian component allocation on different terrains in the survey area, iterate and optimize through the EM algorithm until the model converges. After correcting the characteristics of the collected data as the input vector of the model according to the fuzzy rule base, input it into the model, and output the error probability density representing the terrain error in the survey area to achieve high-precision construction optimization.

2. The intelligent surveying method for the pile foundation pre-embedded area based on fuzzy coordinate compensation according to claim 1, characterized in that: The constraint conditions are the terrain mutation areas where the determined slope value is greater than the slope threshold according to the engineering design requirements, and the key node areas of the pile group with a radius of n times the average pile spacing d around the pile foundation coordinates (x k , y k ); The measurement point layout logic that meets the above constraints is as follows: For the critical node area of the pile group foundation: within the survey area with a radius of nd around the pile foundation coordinates (x k , y k ), the grid density is increased to n times the original, and the grid size is reduced to Δ final / n. Among them, the coordinates of the dense measurement points output are alternately offset based on the starting point according to the parity of the index k, with a horizontal interval of Δ final / n and an increasing fixed step length in the vertical direction; for the terrain mutation area: the coordinates of the dense measurement points output are dynamically adjusted according to the local curvature of the terrain.

3. The intelligent surveying and mapping method for the pre-embedded area of pile foundation based on fuzzy coordinate compensation according to claim 1, wherein: Dynamically adjust the grid size Δ final The specific steps are as follows: First, calculate the measurement point density, ρ(i,j)=ρ base ×(1 + h×W nor (i,j)), where W nor (i,j) is the value of the grid cell (i,j) in the normalized constraint correlation matrix; Secondly, through the formula , calculate the terrain complexity estimate C. In the formula, ρ base is the basic measurement point density, ρ base = 1 / Δ 2 , which represents the expected number of basic measurement points within the unit grid area. The coefficient h is the amplification coefficient of the constraint intensity on the measurement point density, Δ is the basic grid size, Δ = d / (k×C), k is the adjustment coefficient, is the slope value of the i-th measurement point, is the curvature value of the i-th measurement point, is the maximum slope value within the surveyed area, is the maximum curvature value within the surveyed area, and a and b are the weight coefficients of slope and curvature respectively; Finally, through the formula , the grid size Δ final is dynamically adjusted, where η is an adjustment index and ξ is an adjustment coefficient.

4. The intelligent surveying and mapping method for the pre-embedded area of pile foundations based on fuzzy coordinate compensation according to claim 1, characterized in that: The specific steps for constructing a structured data set are: First, dynamically adjust the grid size according to the terrain complexity and pile foundation layout; Secondly, perform coordinate mapping to map the coordinates of the measurement points to the grid indices (i, j): , , where and are the minimum coordinates of the survey area; Finally, perform Z-score normalization on the elevation data to construct a dataset D that contains the normalized elevation data Z' i,j and the corresponding point cloud file PC xc , yc .las, 5. The intelligent surveying and mapping method for the pre-embedded area of pile foundation based on fuzzy coordinate compensation according to claim 1 or 4, characterized in that: The specific steps for using fuzzy clustering to extract the characteristics of the collected data are: First, in the structured dataset D, extract the normalized elevation value Z' of each grid i,j , the point cloud density N i,j , the elevation deviation e i,j， and convert them into fuzzy sets to generate a fuzzy vector F i ; Secondly, construct a clustering algorithm to minimize the objective function: , where N is the total number of grids, C is the number of clusters, and u ic is the membership degree of grid i to cluster c, indicating the degree to which grid i belongs to cluster c, satisfying ∑ c=1 C u ic = 1, m = 2 is the fuzzy index, vc is the center vector of cluster c, and ||·|| is the Euclidean distance; Again, construct the membership update formula to update the cluster center and realize the recognition of the characteristics of the collected data. Among them, the membership update formula is as follows: , where v d is the center vector of cluster d; Finally, the membership matrix U is obtained: U = [u ic N×C , to represent the membership of each grid to each cluster.​ 6. The intelligent surveying and mapping method for the pre-embedded area of pile foundation based on fuzzy coordinate compensation according to claim 5, characterized in that: During the process of updating the cluster center, when the number of iterations reaches Tmax = 100, stop the iteration regardless of whether the clustering is completely convergent, or stop when any one of the conditions of membership matrix change, cluster center change, or objective function change is met: Membership matrix change condition: The change in the membership matrix U in two consecutive iterations is less than the system-set threshold; Cluster center change condition: The change in the cluster center vc in two consecutive iterations is less than the system-set threshold; Objective function change condition: The change in the objective function J in two consecutive iterations is less than the system-set threshold.

7. The intelligent surveying and mapping method for the pre-embedded area of pile foundation based on fuzzy coordinate compensation according to claim 5, characterized in that: During the process of realizing the recognition of the characteristics of the collected data, it is also necessary to use the method of evaluating the importance of feature clusters to quantify the influence of the key features of each collected data on the terrain error, ensure that the measuring points can accurately cover the terrain mutation area and the key nodes of the pile foundation, and at the same time avoid data redundancy in non-critical areas.

8. The intelligent surveying and mapping method for the pre-embedded area of pile foundations based on fuzzy coordinate compensation according to claim 1 or 7, characterized in that: The specific steps for realizing the correction of the characteristics of the collected data are: First, define the relationship between feature combinations and correction weights based on the fuzzy rule base, and set the rule trigger intensity to be determined by the membership degree of relevant features; secondly, use the mamdani fuzzy controller to infer and construct the output fuzzy set, and obtain the specific correction weight through the centroid method of defuzzification; finally, use the correction weight to adjust the elevation and point cloud density to ensure that the model output is more accurate and adapts to different terrain conditions.

9. The intelligent surveying and mapping method for the pre-embedded area of pile foundation based on fuzzy coordinate compensation according to claim 1, wherein: The calculation formula for the error probability density characterizing the terrain error of the mapping area is as follows: , where M is the number of Gaussian components in the model. The model consists of M Gaussian components, and the weight of each component is ωj, the mean is , and the variance is , is the Gaussian distribution density function, representing the probability density of the error e under the j-th Gaussian component, is the mean of the j-th Gaussian component, representing the central position of this component, is the variance of the j-th Gaussian component, representing the distribution range of this component.

10. An intelligent surveying and mapping device for the pre-embedded area of pile foundations based on fuzzy coordinate compensation, characterized in that: including: a memory and a processor, wherein the memory includes an intelligent surveying program for the pile foundation pre-embedded area based on fuzzy coordinate compensation, and when the intelligent surveying program for the pile foundation pre-embedded area based on fuzzy coordinate compensation is executed by the processor, it realizes the intelligent surveying method for the pile foundation pre-embedded area based on fuzzy coordinate compensation according to any one of claims 1 to 9.

Images