Hydraulic engineering surveying and mapping data acquisition system and acquisition method

Abstract

The invention discloses a hydraulic engineering surveying and mapping data acquisition system and an acquisition method, and belongs to the field of hydraulic engineering surveying and mapping data acquisition. Through signal synchronous fusion, clustering point cloud construction, boundary accurate extraction, neural network continuity discrimination, water body space relation calibration and iterative optimization, an enhanced structure model is finally generated and is fed back and updated in real time, and high-precision and high-reliability hydraulic engineering surveying and mapping data acquisition in a complex water area environment is realized. And spatial data support is provided for water level simulation and engineering safety monitoring.

Description

Technical Field

[0001] This invention belongs to the field of water conservancy engineering surveying and mapping data acquisition, and particularly relates to a water conservancy engineering surveying and mapping data acquisition system and acquisition method. Background Technology

[0002] Surveying and mapping of water conservancy projects is a core component in ensuring the safe construction and operation of large dams, waterways, and irrigation systems. The accuracy and real-time nature of its data collection directly affect project decision-making and risk prevention and control, playing a crucial role in national infrastructure construction and flood control and disaster reduction.

[0003] Current traditional surveying methods rely on manual interpretation and fixed rule processing. Their main drawback is that they cannot effectively integrate the dynamic changes of multiple instrument signals, which weakens the correlation of data in complex on-site environments. Especially under conditions of variable lighting or water flow interference, the collected coordinate information is prone to deviation and cannot fully capture the true spatial distribution of the engineering structure.

[0004] At water conservancy project sites, the distance, angle, and reflection intensity values ​​returned by surveying instruments constitute the core elements of data acquisition. These elements are closely interdependent: distance and angle values ​​define the spatial location of the point cloud, while reflection intensity values ​​reflect the differences in surface materials. This dependence poses a primary challenge to data processing: how to accurately correlate these multidimensional signals during real-time acquisition to construct a complete understanding of the spatial structure. If the correlation is improper, subtle fluctuations in reflection intensity values ​​will be difficult to match with location information, leading to a chain of errors in subsequent identification.

[0005] For example, in dam surveying, when the instrument scans the boundary between the bedrock and soil layers, the reflection intensity changes drastically due to the material differences. However, if the distance and angle values ​​are not precisely synchronized and calibrated, the boundary position will be misjudged, causing a fault in the virtual model of the entire dam structure, which in turn affects the reliability of water level simulation. Therefore, how to accurately determine the spatial relationships between terrain features, engineering boundaries, and water distribution based on the dynamic correlation characteristics of these instrument signals when surveying data flows in in real time has become a key issue in improving the efficiency and reliability of water conservancy project data acquisition. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the present invention provides a water conservancy engineering surveying data acquisition system and acquisition method, comprising: The signal synchronization processing module is used to process dynamic changes by collecting distance values, angle values ​​and reflection intensity values, and obtain a fused signal to reflect the differences in surface material. The clustering and grouping module is used to apply a clustering algorithm to group similar features in a complex environment based on the fused signal, and to determine the data association strength in order to construct a preliminary point cloud. The boundary judgment module is used to perform boundary judgment on the preliminary point cloud if the data association strength exceeds a preset threshold, so as to obtain the accurate coordinate distribution of the engineering boundary. The neural network analysis module is used to analyze the correspondence between spatial location and reflection intensity using neural network analysis based on the obtained precise coordinate distribution, and to determine the continuity of terrain features; The deviation calibration module is used to obtain the input data of the structural model based on the determined continuity of the terrain features, and to determine the spatial relationship of water body distribution in order to calibrate the deviation. The iterative optimization module is used to update the point cloud construction through iterative optimization if the spatial relationship after deviation calibration matches the initial fused signal, thereby obtaining an enhanced structural model. The real-time feedback module is used to process subsequent signals in real-time based on the enhanced structural model to obtain the final spatial understanding to support water level simulation and complete the acquisition of mapping data.

[0007] Optionally, the signal synchronization processing module includes: The raw data acquisition unit is used to collect distance, angle, and reflection intensity values ​​through sensors to obtain the raw dataset. A dynamic synchronization unit is used to process the dynamic changes of the original dataset using a signal synchronization mechanism to obtain a synchronized dataset. An intensity distribution extraction unit is used to extract reflection intensity features based on the synchronous dataset and determine the intensity distribution; An intensity grouping unit is used to group the intensity distribution using a K-means clustering algorithm if the intensity distribution matches a preset threshold, and obtain the grouping results. A signal fusion unit is used to fuse distance and angle values ​​based on the grouping results to obtain a fused signal; The material difference judgment unit is used to compare the surface material parameters through the fused signal to judge the material difference; The dataset update unit is used to update the synchronized dataset based on the results of judging material differences, thereby obtaining the optimized dataset.

[0008] Optionally, the clustering grouping module includes: The initial data acquisition unit is used to acquire an initial data set through the fused signal and determine the distribution of similar features; K-means clustering units are used to group data points according to the distribution of similar features using the K-means clustering algorithm to obtain the grouping results; The association strength calculation unit is used to calculate the Pearson correlation coefficient based on the grouping results and determine the association strength value. The low-intensity filtering unit is used to filter low-intensity associations based on the association intensity value to obtain a filtered data set. The noise removal unit is used to process the filtered dataset using the DBSCAN algorithm, identify and remove noise points, and obtain a refined dataset. The 3D mapping unit is used to construct a 3D coordinate mapping through a refined dataset and determine the initial point cloud structure. The coordinate offset adjustment unit is used to adjust the coordinate offset according to the initial point cloud structure to obtain an optimized point cloud.

[0009] Optionally, the boundary determination module includes: The threshold comparison unit is used to compare the correlation strength with a preset threshold. If the threshold is exceeded, feature points are extracted from the point cloud data to obtain a boundary candidate set. The clustering cluster determination unit is used to group the point cloud according to the boundary candidate set using a clustering algorithm to determine the clusters of the point cloud; The cluster merging unit is used to calculate spatial distance by clustering clusters. If the distance is lower than a threshold, the clusters are merged to obtain an optimized boundary set. The curvature analysis unit is used to perform curvature analysis on the optimized boundary set, identify curvature abrupt change points, and obtain the boundary turning points. Interpolation connection unit, used to interpolate and connect based on the boundary turning points to determine continuous boundary lines; The coordinate distribution output unit is used to obtain the accurate coordinate distribution of the engineering boundary by mapping continuous boundary lines to a coordinate system.

[0010] Optionally, the neural network analysis module includes: The distribution data extraction unit is used to obtain precise coordinate distribution and extract spatial location data and reflection intensity data from the distribution. The neural network processing unit is used to process spatial location data and reflection intensity data using a neural network to obtain a corresponding relationship matrix. The continuity index calculation unit is used to calculate the continuity index through the correspondence matrix and determine the terrain feature sequence; The continuity judgment unit is used to determine the continuity if the difference between adjacent elements in the terrain feature sequence is less than a preset threshold, and thus obtain a continuous segment. The region boundary division unit is used to divide the region boundary according to continuous segments and obtain the boundary coordinate set; The overall continuity verification unit is used to determine the overall terrain continuity by mapping the boundary coordinate set to the original coordinate distribution. Boundary coordinate clustering units are used to group boundary coordinate sets using clustering algorithms to obtain grouped clusters.

[0011] Optionally, the deviation calibration module includes: The remote sensing image acquisition unit is used to acquire terrain features through remote sensing images, determine continuity, and obtain continuous segments; The clustering input determination unit is used to process the continuous paragraphs using a clustering algorithm and determine the structural model input data if the number of consecutive paragraphs exceeds a preset threshold. Mesh building units are used to construct a mesh based on input data from the structural model, thereby obtaining the water body distribution and its boundaries. The adjacency analysis unit is used to analyze adjacency points through distributed boundaries, determine spatial relationships, and obtain a relationship matrix; The random forest calibration unit is used to calibrate the relation matrix by fusing the random forest algorithm if the element value in the relation matrix is ​​greater than zero, and to obtain the bias vector. The distribution boundary adjustment unit is used to adjust the distribution boundary using the deviation vector, determine the calibration deviation, and obtain the calibration result.

[0012] Optionally, the iterative optimization module includes: The matching degree index extraction unit is used to process spatial relationship data through deviation calibration, obtain the matching degree index from the fused signal, and obtain the deviation calibration result; Point cloud grouping unit is used to group point cloud data using K-means clustering algorithm if the deviation calibration result shows that the spatial relationship matches the initial fusion signal, so as to obtain a grouped point cloud set; The vector similarity calculation unit is used to calculate the spatial relationship vector based on the grouped point cloud set, determine the vector similarity, and obtain the basis for similarity threshold judgment. The iterative optimization unit is used to adjust the point cloud position through an iterative optimization algorithm to obtain an optimized point cloud framework if the similarity threshold judgment criteria exceed a preset threshold. The feature consistency verification unit is used to fuse the features of the initial fused signal using an optimized point cloud framework, judge the feature consistency, and obtain the consistency verification output. The model draft generation unit is used to update the point cloud construction parameters through consistency verification to obtain the enhanced structural model draft. The surface curve fitting unit is used to fit the surface curves according to the draft of the reinforced structure model using the least squares method to obtain the final reinforced structure model.

[0013] On the other hand, the present invention also provides a method for acquiring surveying and mapping data for water conservancy projects, comprising: The distance, angle, and reflection intensity values ​​collected simultaneously are synchronized and weighted to obtain a fused signal; Based on the fused signal, similar features are grouped and data association strength is calculated using a clustering algorithm in a complex environment, and a preliminary point cloud is constructed based on the data association strength. When the data association strength exceeds a preset threshold, a boundary judgment is performed on the preliminary point cloud to obtain the precise coordinate distribution of the engineering boundary; The precise coordinate distribution is input into a neural network to analyze the correspondence between spatial location and reflection intensity, and output the result of the terrain feature continuity judgment. Based on the results of the terrain feature continuity judgment, structural model input data is obtained, the spatial relationship of water body distribution is determined, and the spatial relationship is calibrated for deviation. If the spatial relationship after deviation calibration matches the fused signal, the point cloud is updated through iterative optimization to generate an enhanced structural model. Using the enhanced structural model as a reference, the subsequent acquired signals are processed in real time feedback loop to obtain the final spatial understanding data for water level simulation.

[0014] On the other hand, the present invention also provides an electronic device including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computing program.

[0015] On the other hand, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0016] Compared with the prior art, the present invention has the following advantages and technical effects: To address the issues of inaccurate point cloud data association and boundary judgment deviations caused by surface material differences in dynamically changing scenarios, this invention addresses a chain of factors including signal interference, unstable feature clustering, and difficulties in determining spatial continuity, all of which negatively impact water level simulation accuracy. This invention collects distance, angle, and reflection intensity values, employs a signal synchronization mechanism to generate a fused signal reflecting material differences, and then applies a clustering algorithm to group similar features in complex environments, calculating data association strength to construct an initial point cloud. When the association strength exceeds a threshold, boundary judgment is performed to obtain precise coordinate distribution. A neural network is then used to analyze the correspondence between coordinates and reflection intensity to determine terrain continuity. Based on this, a structural model input is obtained to determine the spatial relationship of water body distribution and calibrate deviations. If the calibrated relationship matches the initial signal, the point cloud is iteratively optimized to construct an enhanced structural model. Real-time feedback is used to iteratively process subsequent signals, ultimately achieving accurate spatial understanding to support water level simulation. The core analytical business scenario of this invention—the provided text describes a process for processing multi-source sensor data, including distance, angle, and reflection intensity, for point cloud construction and terrain analysis in complex environments. The key lies in the iterative calibration mechanism driven by the fusion of signals and neural networks, which ensures the robustness and boundary accuracy of point clouds in dynamic environments. The overall technical effect is to improve the accuracy of water body topography modeling, which is suitable for engineering monitoring and flood prediction scenarios. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of the system structure according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the adjustment boundary according to an embodiment of the present invention. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0020] Example 1 like Figure 1 As shown, this embodiment provides a water conservancy engineering surveying and mapping data acquisition system, including: The signal synchronization processing module is used to process dynamic changes by collecting distance values, angle values ​​and reflection intensity values, and obtain a fused signal to reflect the differences in surface material. The clustering and grouping module is used to apply a clustering algorithm to group similar features in a complex environment based on the fused signal, and to determine the data association strength in order to construct a preliminary point cloud. The boundary judgment module is used to perform boundary judgment on the preliminary point cloud if the data association strength exceeds a preset threshold, so as to obtain the accurate coordinate distribution of the engineering boundary. The neural network analysis module is used to analyze the correspondence between spatial location and reflection intensity using neural network analysis based on the obtained precise coordinate distribution, and to determine the continuity of terrain features; The deviation calibration module is used to obtain the input data of the structural model based on the determined continuity of the terrain features, and to determine the spatial relationship of water body distribution in order to calibrate the deviation. The iterative optimization module is used to update the point cloud construction through iterative optimization if the spatial relationship after deviation calibration matches the initial fused signal, thereby obtaining an enhanced structural model. The real-time feedback module is used to process subsequent signals in real-time based on the enhanced structural model to obtain the final spatial understanding to support water level simulation and complete the acquisition of mapping data.

[0021] Optionally, the signal synchronization processing module includes: The raw data acquisition unit is used to collect distance, angle, and reflection intensity values ​​through sensors to obtain the raw dataset. A dynamic synchronization unit is used to process the dynamic changes of the original dataset using a signal synchronization mechanism to obtain a synchronized dataset. An intensity distribution extraction unit is used to extract reflection intensity features based on the synchronous dataset and determine the intensity distribution; An intensity grouping unit is used to group the intensity distribution using a K-means clustering algorithm if the intensity distribution matches a preset threshold, and obtain the grouping results. A signal fusion unit is used to fuse distance and angle values ​​based on the grouping results to obtain a fused signal; The material difference judgment unit is used to compare the surface material parameters through the fused signal to judge the material difference; The dataset update unit is used to update the synchronized dataset based on the results of judging material differences, thereby obtaining the optimized dataset.

[0022] Specifically, by collecting distance, angle, and reflection intensity values, a signal synchronization mechanism is used to process dynamic changes and obtain a fused signal to reflect differences in surface material.

[0023] A raw dataset is obtained by collecting distance, angle, and reflection intensity values ​​using sensors. A signal synchronization mechanism is then used to process the dynamic changes in the raw dataset, resulting in a synchronized dataset. Reflection intensity features are extracted from the synchronized dataset to determine the intensity distribution. If the intensity distribution matches a preset threshold, it is grouped using a K-means clustering algorithm to obtain grouping results. Distance and angle values ​​are then fused based on the grouping results to obtain a fused signal. The fused signal is compared with surface material parameters to determine material differences. The synchronized dataset is then updated based on the determined material differences to obtain an optimized dataset.

[0024] Specifically, the system first uses a lidar sensor to collect data in the scanning environment. For example, at timestamp t=0.1 seconds, it collects distance values ​​d=2.5 meters, angle values ​​θ=30 degrees, and reflection intensity values ​​i=80. Then, a multi-sensor array automatically collects multiple sets of data, such as d=2.6 meters, θ=31 degrees, and i=82 at t=0.2 seconds, to form a time-series dataset. These data are then processed using preprocessing algorithms such as median filtering to remove noise. For example, the distance series is filtered using the formula d_filtered=median(d_t-1,d_t,d_t+1) to calculate a smooth value of 2.55 meters, thus ensuring data accuracy and providing a basis for subsequent synchronization. Next, a signal synchronization mechanism is used to handle dynamic changes. For example, a clock synchronization protocol such as the PTP algorithm is used to align data from different sensors to a unified time base. For dynamic changes such as distance fluctuations caused by object movement, a Kalman filter algorithm is used to fuse predictions. For example, the state equation is x_k=Fx_{k-1}+w_k, where F is the transition matrix set to [1,0.1;0,1], and the observation equation is z_k=Hx_k+v_k with H=[1,0]. Through iterative calculation of posterior estimation, such as predicting from an initial distance of 2.5 meters to 2.52 meters, the analysis process shows that the variance after filtering decreases from 0.05 to 0.02, indicating that dynamic noise caused by environmental vibration is effectively suppressed. It is also related to business operations, such as adjusting the path in real time in robot navigation to avoid collisions, forming a logical chain from data acquisition to stable signals. The final fused signal reflects the differences in surface materials. For example, the fused value s = 0.4*d + 0.3*θ + 0.3*i is calculated using a weighted fusion algorithm. For metal surfaces, s = 1.8 is obtained, while for plastic surfaces, s = 1.2. Analysis shows that the difference is dominated by the reflection intensity. For example, an intensity threshold > 70 indicates a smooth material. A machine learning model, such as SVM, is used to classify and train the dataset containing 100 samples with an accuracy of 95%. This generates a fused signal that reflects the material for subsequent applications, such as automatic classification systems, ensuring a rigorous logical relationship from acquisition to fusion.

[0025] Optionally, the clustering grouping module includes: The initial data acquisition unit is used to acquire an initial data set through the fused signal and determine the distribution of similar features; K-means clustering units are used to group data points according to the distribution of similar features using the K-means clustering algorithm to obtain the grouping results; The association strength calculation unit is used to calculate the Pearson correlation coefficient based on the grouping results and determine the association strength value. The low-intensity filtering unit is used to filter low-intensity associations based on the association intensity value to obtain a filtered data set. The noise removal unit is used to process the filtered dataset using the DBSCAN algorithm, identify and remove noise points, and obtain a refined dataset. The 3D mapping unit is used to construct a 3D coordinate mapping through a refined dataset and determine the initial point cloud structure. The coordinate offset adjustment unit is used to adjust the coordinate offset according to the initial point cloud structure to obtain an optimized point cloud.

[0026] Specifically, based on the fused signal, a clustering algorithm is applied in a complex environment to group similar features and determine the data association strength in order to construct a preliminary point cloud.

[0027] An initial dataset is obtained by fusing signals to determine the distribution of similar features. Data points are then grouped using the K-means clustering algorithm based on this distribution. The Pearson correlation coefficient is calculated using these grouping results to determine the association strength. Low-strength associations are filtered out based on the association strength, resulting in a filtered dataset. The DBSCAN algorithm is then used to process the filtered dataset, identifying and removing noise points to obtain a refined dataset. A 3D coordinate mapping is constructed using the refined dataset to determine the initial point cloud structure. The coordinate offset is adjusted based on the initial point cloud structure to obtain an optimized point cloud.

[0028] Specifically, firstly, a Kalman filter algorithm is used to fuse multi-source signal data from LiDAR and cameras. For example, the distance measurement values ​​from LiDAR and the image pixels from the camera are weighted and averaged. Assuming the LiDAR data weight is 0.7 and the camera weight is 0.3, a unified fused signal dataset is generated. Each data point contains coordinates (x, y, z) and an intensity value, such as an average intensity of 85.2, to reduce noise interference and improve data accuracy. Next, in complex environments such as urban traffic scenarios, a K-means clustering algorithm is applied to group similar features of these fused signals. The specific process includes initializing K=5 cluster centers and iteratively calculating the Euclidean distance. For example, clustering is performed on feature vectors (intensity 85.2, RGB color values ​​[120, 150, 200]). When the distance threshold is less than 2.5, they are grouped into the same cluster. Through analysis, it was found that the intra-cluster variance decreased from an initial 15.6 to a final 3.2. This indicates that the grouping is effective. This is combined with business-related factors such as traffic flow monitoring to optimize cluster stability. Then, the data association strength is determined using the Pearson correlation coefficient. For example, a coefficient of 0.85 for data points in two clusters indicates a strong positive correlation. Through reasoning, if the coefficient is greater than 0.8, it is considered a high correlation, thus linking related feature points to form a network structure and avoiding isolated points from affecting the overall model. Finally, based on these association strengths, a preliminary point cloud is constructed, and a 3D mesh is generated using the Delaunay triangulation algorithm. For example, after inputting the associated point set, the total number of triangular faces is calculated to be 1200, and interpolation is used to fill gaps to ensure that the point cloud density reaches 500 points per cubic meter. The analysis process verifies that the convex hull volume of the point cloud is 150.4 cubic meters, which meets the initial reconstruction requirements in the expected complex environment. The entire process is logically from signal fusion to provide basic data, to clustering and grouping to extract patterns, to quantifying and linking association strengths, and finally constructing a point cloud to form a visual output.

[0029] Optionally, the boundary determination module includes: The threshold comparison unit is used to compare the correlation strength with a preset threshold. If the threshold is exceeded, feature points are extracted from the point cloud data to obtain a boundary candidate set. The clustering cluster determination unit is used to group the point cloud according to the boundary candidate set using a clustering algorithm to determine the clusters of the point cloud; The cluster merging unit is used to calculate spatial distance by clustering clusters. If the distance is lower than a threshold, the clusters are merged to obtain an optimized boundary set. The curvature analysis unit is used to perform curvature analysis on the optimized boundary set, identify curvature abrupt change points, and obtain the boundary turning points. Interpolation connection unit, used to interpolate and connect based on the boundary turning points to determine continuous boundary lines; The coordinate distribution output unit is used to obtain the accurate coordinate distribution of the engineering boundary by mapping continuous boundary lines to a coordinate system.

[0030] Specifically, if the data association strength exceeds a preset threshold, boundary judgment is performed on the point cloud to obtain the precise coordinate distribution of the engineering boundary.

[0031] By comparing the correlation strength with a preset threshold, if the threshold is exceeded, feature points are extracted from the point cloud data to obtain a candidate set of boundaries. A clustering algorithm is then used to group the point cloud data into clusters. Spatial distances are calculated between these clusters; if the distance is below a threshold, the clusters are merged to obtain an optimized boundary set. Curvature analysis is performed on the optimized boundary set to identify curvature abrupt changes and determine boundary inflection points. Interpolation connections are then made based on these boundary inflection points to determine continuous boundary lines. Finally, these continuous boundary lines are mapped to a coordinate system to obtain the precise coordinate distribution of the engineering boundary.

[0032] Specifically, when processing point cloud data, the Pearson correlation coefficient is first used to assess the strength of the data association. For example, analyzing the coordinate vectors of two point cloud datasets, assuming dataset A contains 1000 point coordinates such as (1.2, 3.4, 5.6), and dataset B contains the corresponding points (1.3, 3.5, 5.7), the correlation coefficient formula is r=∑(xi-¯x)(yi-¯y) / √[∑(xi-¯x)] 2 ∑(yi-¯y) 2 The result is r=0.95. If the preset threshold is 0.8, then 0.95>0.8 satisfies the condition, triggering subsequent processing to ensure data reliability and continuity of engineering applications. Next, boundary judgment is performed on the point cloud. The AlphaShapes algorithm is used to generate boundary polygons. Specifically, this involves constructing a Delaunay triangulation network, removing triangle edges with radii greater than alpha (e.g., 10.0), analyzing the remaining connections to form closed boundary curves, and extracting boundary point sets from the point cloud, reducing the total number of points to 500. Boundary smoothness is verified through curvature calculation, with an average curvature of 0.05, logically connecting to precise coordinate acquisition to support downstream engineering modeling. Finally, the precise coordinate distribution of the engineering boundary is obtained, and the boundary point coordinates are optimized using a least-squares fitting algorithm, for example, applying a polynomial fit y=ax to the boundary curve. 2 The algorithm +bx+c solves for parameters a=0.1, b=2.0, and c=3.0, generating distribution sequences such as (0.0,3.0), (1.0,5.1), and (2.0,9.4). Error analysis is performed, and the root mean square error is 0.02, ensuring that the coordinate accuracy reaches the millimeter level for integration into the CAD system.

[0033] Optionally, the neural network analysis module includes: The distribution data extraction unit is used to obtain precise coordinate distribution and extract spatial location data and reflection intensity data from the distribution. The neural network processing unit is used to process spatial location data and reflection intensity data using a neural network to obtain a corresponding relationship matrix. The continuity index calculation unit is used to calculate the continuity index through the correspondence matrix and determine the terrain feature sequence; The continuity judgment unit is used to determine the continuity if the difference between adjacent elements in the terrain feature sequence is less than a preset threshold, and thus obtain a continuous segment. The region boundary division unit is used to divide the region boundary according to continuous segments and obtain the boundary coordinate set; The overall continuity verification unit is used to determine the overall terrain continuity by mapping the boundary coordinate set to the original coordinate distribution. Boundary coordinate clustering units are used to group boundary coordinate sets using clustering algorithms to obtain grouped clusters.

[0034] Specifically, by using the obtained precise coordinate distribution, a neural network is employed to analyze the correspondence between spatial location and reflection intensity, thereby determining the continuity of terrain features.

[0035] A precise coordinate distribution is obtained, and spatial location data and reflectance intensity data are extracted from this distribution. A neural network is used to process the spatial location data and reflectance intensity data to obtain a correspondence matrix. A continuity index is calculated using this correspondence matrix to determine the terrain feature sequence. If the difference between adjacent elements in the terrain feature sequence is less than a preset threshold, continuity is determined, and continuous segments are obtained. The region boundaries are divided based on these continuous segments to obtain boundary coordinate sets. The overall terrain continuity is determined by mapping the boundary coordinate sets back to the original coordinate distribution. A clustering algorithm is used to group the boundary coordinate sets, resulting in clusters.

[0036] Specifically, the first step is to use a lidar system to scan the terrain area and obtain precise coordinate distribution data. For example, 10,000 points are collected within a 100m × 100m grid. Each point records its three-dimensional coordinates (x=50.2, y=30.5, z=10.8) and corresponding reflection intensity value (e.g., 0.75). The data is then imported into a DataFrame structure using Python's Pandas library for preprocessing, including normalizing the coordinates to the [0,1] range and filtering out noisy points. This ensures that the data quality reaches an accuracy of over 95%, thus providing a reliable foundation for subsequent analysis. Next, a multilayer perceptron neural network model was adopted. The input layer received the spatial location vector (x, y, z) and reflection intensity as 4-dimensional features. The hidden layer consisted of two layers with 128 neurons each, using the ReLU activation function. The output layer predicted the continuity score. The model was trained using the Adam optimizer with a learning rate of 0.001. The dataset was divided into an 80% training set and a 20% test set. The training process iterated for 100 epochs, and the loss function was calculated as mean squared error. During the analysis, the correspondence between reflection intensity and location was observed. For example, an intensity of 0.75 at coordinates (x=50.2, y=30.5) corresponds to flat terrain, while the intensity of 0.72 at the neighboring point (x=50.3, y=30.6) shows a continuous transition. The model minimized the error to below 0.05 through gradient descent, revealing that the region with an intensity change rate of less than 0.1 is a continuous feature. Finally, the continuity of terrain features is judged based on the output of the neural network. If the prediction score exceeds the 0.8 threshold, it is considered continuous. For example, the proportion of clusters with an Euclidean distance of less than 1 meter between adjacent points and an intensity difference of less than 0.05 is calculated to reach 70%. The continuity is verified by combining business correlations such as integrating satellite image data.

[0037] Optionally, the deviation calibration module includes: The remote sensing image acquisition unit is used to acquire terrain features through remote sensing images, determine continuity, and obtain continuous segments; The clustering input determination unit is used to process the continuous paragraphs using a clustering algorithm and determine the structural model input data if the number of consecutive paragraphs exceeds a preset threshold. Mesh building units are used to construct a mesh based on input data from the structural model, thereby obtaining the water body distribution and its boundaries. The adjacency analysis unit is used to analyze adjacency points through distributed boundaries, determine spatial relationships, and obtain a relationship matrix; The random forest calibration unit is used to calibrate the relation matrix by fusing the random forest algorithm if the element value in the relation matrix is ​​greater than zero, and to obtain the bias vector. The distribution boundary adjustment unit is used to adjust the distribution boundary using the deviation vector, determine the calibration deviation, and obtain the calibration result.

[0038] Based on the determined continuity of topographic features, the input data of the structural model is obtained, and the spatial relationship of water body distribution is determined to calibrate the deviation.

[0039] Topographic features are acquired through remote sensing images, and continuity is determined to identify continuous segments. If the number of continuous segments exceeds a preset threshold, a clustering algorithm is used to process these segments and determine the input data for the structural model. A grid is constructed based on the structural model input data to obtain the water body distribution and its boundaries. Adjacent points are analyzed through the distribution boundaries to determine spatial relationships and obtain a relation matrix. If any element in the relation matrix is ​​greater than zero, a random forest algorithm is used to calibrate the relation matrix, resulting in a deviation vector. The deviation vector is then used to adjust the distribution boundaries, and the calibration deviation is assessed to obtain the calibration result.

[0040] Specifically, Geographic Information System (GIS) software is used to determine the continuity of terrain features in satellite remote sensing images. For example, a slope gradient continuity algorithm is used, with the image resolution set to 0.5 meters. The analysis process includes extracting areas from the DEM data with elevation values ​​ranging from 100 to 500 meters, applying Moran's I autocorrelation algorithm to calculate a continuity index. If the index is greater than 0.8, it is considered continuous. This index reflects the smoothness of the terrain to avoid misjudging faults, forming a thought chain connecting to subsequent data acquisition, as the continuity determination result directly affects the reliability of the model input. Based on the determination result, structural model input data is automatically acquired. For example, a Python script is used to extract the coordinate point set of continuous terrain areas from the database, including latitude and longitude values ​​such as 30.5 degrees North latitude and 120.3 degrees East longitude, and integrate soil permeability data of 0.2 cm / hour. A three-dimensional mesh model is generated using the Kriging interpolation algorithm, and the analysis is then performed. The process involves error minimization calculations to ensure data accuracy of over 95%. This step is linked to other business operations, such as flood simulation, to expand applications and form a thought chain to ensure that input data supports water body distribution analysis. Then, the spatial relationships of water body distribution are determined. For example, ArcGIS spatial analysis tools are used to calculate the buffer zones between river networks and lakes, setting the buffer distance to 50 meters. The algorithm uses the Euclidean distance matrix to calculate a spatial overlap rate of 70%. The analysis process includes classifying water body types and quantifying the distribution density to 2.5 per square kilometer. This relationship is explained to identify potential floodplains to calibrate model bias. Finally, the bias is calibrated through iterative optimization algorithms. For example, the least squares method is applied to compare simulated water body distribution with actual remote sensing data, adjusting parameters to reduce the bias rate from the initial 15% to 2.5%. The analysis process involves 100 iterations of cyclic verification to converge, forming a thought chain that connects back to terrain continuity judgment to improve overall model accuracy.

[0041] Optionally, the iterative optimization module includes: The matching degree index extraction unit is used to process spatial relationship data through deviation calibration, obtain the matching degree index from the fused signal, and obtain the deviation calibration result; Point cloud grouping unit is used to group point cloud data using K-means clustering algorithm if the deviation calibration result shows that the spatial relationship matches the initial fusion signal, so as to obtain a grouped point cloud set; The vector similarity calculation unit is used to calculate the spatial relationship vector based on the grouped point cloud set, determine the vector similarity, and obtain the basis for similarity threshold judgment. The iterative optimization unit is used to adjust the point cloud position through an iterative optimization algorithm to obtain an optimized point cloud framework if the similarity threshold judgment criteria exceed a preset threshold. The feature consistency verification unit is used to fuse the features of the initial fused signal using an optimized point cloud framework, judge the feature consistency, and obtain the consistency verification output. The model draft generation unit is used to update the point cloud construction parameters through consistency verification to obtain the enhanced structural model draft. The surface curve fitting unit is used to fit the surface curves according to the draft of the reinforced structure model using the least squares method to obtain the final reinforced structure model.

[0042] If the spatial relationship after deviation calibration matches the initial fused signal, then the enhanced structural model is obtained by iteratively optimizing and updating the point cloud.

[0043] Spatial relationship data is processed by deviation calibration, and a matching degree index is obtained from the fused signal to obtain the deviation calibration result. If the deviation calibration result shows that the spatial relationship matches the initial fused signal, the point cloud data is grouped using the K-means clustering algorithm to obtain a grouped point cloud set. Based on the grouped point cloud set, the spatial relationship vector is calculated, the vector similarity is determined, and the similarity threshold judgment criterion is obtained. If the similarity threshold judgment criterion exceeds the preset threshold, the point cloud position is adjusted through an iterative optimization algorithm to obtain an optimized point cloud framework. Using the obtained optimized point cloud framework, the features of the initial fused signal are fused, the feature consistency is judged, and a consistency verification output is obtained. Based on the consistency verification output, the point cloud construction parameters are updated to obtain a draft enhanced structure model. Based on the draft enhanced structure model, the least squares method is applied to fit the surface curve to obtain the final enhanced structure model.

[0044] Specifically, in the deviation calibration stage, the system first uses the Iterative Closest Point (ICP) algorithm to calibrate the acquired point cloud data. Assuming the initial point cloud contains 10,000 points, the Euclidean distance between each point and the reference model is calculated, with an average deviation of 0.05 meters. The rotation matrix and translation vector are optimized using the least squares method. After five iterations, the deviation is reduced to below 0.01 meters, ensuring a 20% improvement in spatial alignment accuracy. Subsequently, the matching degree between the calibrated spatial relationship and the initial fused signal is analyzed. Specifically, the covariance matrix of the point cloud is calculated and compared with the feature vector of the initial signal. If the matching degree exceeds 95% (based on the cosine similarity formula cosθ=A·), the system is considered to have a better alignment. If B / (|A||B|)), the match is confirmed to be successful; otherwise, the calibration step is returned to form a closed-loop optimization chain. After matching, the iterative optimization and update of the point cloud construction process is entered. The Levenberg-Marquardt algorithm is used to minimize the nonlinearity of the point cloud. The initial value of the damping factor is set to 1.0, and the iteration is performed 10 times. Each time, the point cloud density is updated from 500 points per cubic meter to 1000 points per cubic meter. At the same time, sensor fusion data such as RGB images are incorporated to enhance texture details. The analysis process shows that the error function decreases from the initial 10.2 to 0.5, forming a logical progression from coarse to fine. Finally, the enhanced structural model is constructed through the voxel meshing method.

[0045] The real-time feedback module includes: The signal strength judgment unit is used to obtain initial signal data based on the enhancement structure model, judge the signal strength through a preset threshold, and if the signal strength exceeds the preset threshold, filter the signal data to obtain a filtered signal. The feature vector extraction unit is used to extract feature vectors from the filtered signal through a feedback loop; K-means grouping unit is used to group feature vectors using the K-means clustering algorithm to determine the grouping results; The trend calculation unit is used to calculate the signal change trend based on the grouping results. If the change trend shows an upward trend, the subsequent signals are fused to update the grouping results and obtain updated groups. The coordinate mapping unit is used to obtain the spatial coordinates in the update group and obtain the mapped coordinates by mapping the spatial coordinates through a convolutional neural network. The feedback integration unit is used to integrate feedback data for the mapped coordinates using a real-time acquisition mechanism. If the feedback data matches the mapped coordinates, the coordinate parameters are adjusted to determine the adjusted coordinates. The simulation parameter extraction unit is used to extract simulation parameters from the adjusted coordinates and fit the simulation parameters to obtain the fitted parameters through a linear regression algorithm. The water level model construction unit is used to build a water level model based on the fitted parameters, integrate the fitted parameters through the information processing stage to judge the water level fluctuation and obtain the fluctuation value, and output the final spatial understanding data.

[0046] Based on the enhanced structural model, subsequent signals are processed in real-time acquisition feedback loop to obtain the final spatial understanding to support water level simulation.

[0047] Initial signal data is acquired based on the enhanced structural model. Signal strength is determined using a preset threshold; if the signal strength exceeds the threshold, the signal data is filtered to obtain a filtered signal. Feature vectors are extracted from the filtered signal through a feedback loop, and K-means clustering is used to group the feature vectors to determine the grouping results. The signal change trend is calculated based on the grouping results; if the trend shows an upward trend, subsequent signals are fused to update the grouping results to obtain updated groups. Spatial coordinates from the updated groups are acquired and mapped using a convolutional neural network to obtain mapped coordinates. A real-time acquisition mechanism is used to integrate feedback data for the mapped coordinates; if the feedback data matches the mapped coordinates, the coordinate parameters are adjusted to determine the adjusted coordinates. Simulation parameters are extracted from the adjusted coordinates and fitted using a linear regression algorithm to obtain fitted parameters. A water level model is constructed based on the fitted parameters, and the water level fluctuations are determined through information processing by fusing the fitted parameters to obtain the fluctuation values.

[0048] On the other hand, such as Figure 2 and Figure 3 As shown in the figure, this embodiment also provides a method for collecting surveying and mapping data for water conservancy projects, including: The distance, angle, and reflection intensity values ​​collected simultaneously are synchronized and weighted to obtain a fused signal; Based on the fused signal, similar features are grouped and data association strength is calculated using a clustering algorithm in a complex environment, and a preliminary point cloud is constructed based on the data association strength. When the data association strength exceeds a preset threshold, a boundary judgment is performed on the preliminary point cloud to obtain the precise coordinate distribution of the engineering boundary; The precise coordinate distribution is input into a neural network to analyze the correspondence between spatial location and reflection intensity, and output the result of the terrain feature continuity judgment. Based on the results of the terrain feature continuity judgment, structural model input data is obtained, the spatial relationship of water body distribution is determined, and the spatial relationship is calibrated for deviation. If the spatial relationship after deviation calibration matches the fused signal, the point cloud is updated through iterative optimization to generate an enhanced structural model. Using the enhanced structural model as a reference, the subsequent acquired signals are processed in real time feedback loop to obtain the final spatial understanding data for water level simulation.

[0049] On the other hand, this embodiment also provides an electronic device, including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computing program.

[0050] On the other hand, this embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0051] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A water conservancy project surveying and mapping data acquisition system, characterized in that, include: The signal synchronization processing module is used to process dynamic changes by collecting distance values, angle values ​​and reflection intensity values, and obtain a fused signal to reflect the differences in surface material. The clustering and grouping module is used to apply a clustering algorithm to group similar features in a complex environment based on the fused signal, and to determine the data association strength in order to construct a preliminary point cloud. The boundary judgment module is used to perform boundary judgment on the preliminary point cloud if the data association strength exceeds a preset threshold, so as to obtain the accurate coordinate distribution of the engineering boundary. The neural network analysis module is used to analyze the correspondence between spatial location and reflection intensity using neural network analysis based on the obtained precise coordinate distribution, and to determine the continuity of terrain features; The deviation calibration module is used to obtain the input data of the structural model based on the determined continuity of the terrain features, and to determine the spatial relationship of water body distribution in order to calibrate the deviation. The iterative optimization module is used to update the point cloud construction through iterative optimization if the spatial relationship after deviation calibration matches the initial fused signal, thereby obtaining an enhanced structural model. The real-time feedback module is used to process subsequent signals in real-time based on the enhanced structural model to obtain the final spatial understanding to support water level simulation and complete the acquisition of mapping data.

2. The system according to claim 1, characterized in that, The signal synchronization processing module includes: The raw data acquisition unit is used to collect distance, angle, and reflection intensity values ​​through sensors to obtain the raw dataset. A dynamic synchronization unit is used to process the dynamic changes of the original dataset using a signal synchronization mechanism to obtain a synchronized dataset. An intensity distribution extraction unit is used to extract reflection intensity features based on the synchronous dataset and determine the intensity distribution; An intensity grouping unit is used to group the intensity distribution using a K-means clustering algorithm if the intensity distribution matches a preset threshold, and obtain the grouping results. A signal fusion unit is used to fuse distance and angle values ​​based on the grouping results to obtain a fused signal; The material difference judgment unit is used to compare the surface material parameters through the fused signal to judge the material difference; The dataset update unit is used to update the synchronized dataset based on the results of judging material differences, thereby obtaining the optimized dataset.

3. The system according to claim 1, characterized in that, The clustering grouping module includes: The initial data acquisition unit is used to acquire an initial data set through the fused signal and determine the distribution of similar features; K-means clustering units are used to group data points according to the distribution of similar features using the K-means clustering algorithm to obtain the grouping results; The association strength calculation unit is used to calculate the Pearson correlation coefficient based on the grouping results and determine the association strength value. The low-intensity filtering unit is used to filter low-intensity associations based on the association intensity value to obtain a filtered data set. The noise removal unit is used to process the filtered dataset using the DBSCAN algorithm, identify and remove noise points, and obtain a refined dataset. The 3D mapping unit is used to construct a 3D coordinate mapping through a refined dataset and determine the initial point cloud structure. The coordinate offset adjustment unit is used to adjust the coordinate offset according to the initial point cloud structure to obtain an optimized point cloud.

4. The system according to claim 1, characterized in that, The boundary determination module includes: The threshold comparison unit is used to compare the correlation strength with a preset threshold. If the threshold is exceeded, feature points are extracted from the point cloud data to obtain a boundary candidate set. The clustering cluster determination unit is used to group the point cloud according to the boundary candidate set using a clustering algorithm to determine the clusters of the point cloud; The cluster merging unit is used to calculate spatial distance by clustering clusters. If the distance is lower than a threshold, the clusters are merged to obtain an optimized boundary set. The curvature analysis unit is used to perform curvature analysis on the optimized boundary set, identify curvature abrupt change points, and obtain the boundary turning points. Interpolation connection unit, used to interpolate and connect based on the boundary turning points to determine continuous boundary lines; The coordinate distribution output unit is used to obtain the accurate coordinate distribution of the engineering boundary by mapping continuous boundary lines to a coordinate system.

5. The system according to claim 1, characterized in that, The neural network analysis module includes: The distribution data extraction unit is used to obtain precise coordinate distribution and extract spatial location data and reflection intensity data from the distribution. The neural network processing unit is used to process spatial location data and reflection intensity data using a neural network to obtain a corresponding relationship matrix. The continuity index calculation unit is used to calculate the continuity index through the correspondence matrix and determine the terrain feature sequence; The continuity judgment unit is used to determine the continuity if the difference between adjacent elements in the terrain feature sequence is less than a preset threshold, and thus obtain a continuous segment. The region boundary division unit is used to divide the region boundary according to continuous segments and obtain the boundary coordinate set; The overall continuity verification unit is used to determine the overall terrain continuity by mapping the boundary coordinate set to the original coordinate distribution. Boundary coordinate clustering units are used to group boundary coordinate sets using clustering algorithms to obtain grouped clusters.

6. The system according to claim 1, characterized in that, The deviation calibration module includes: The remote sensing image acquisition unit is used to acquire terrain features through remote sensing images, determine continuity, and obtain continuous segments; The clustering input determination unit is used to process the continuous paragraphs using a clustering algorithm and determine the structural model input data if the number of consecutive paragraphs exceeds a preset threshold. Mesh building units are used to construct a mesh based on input data from the structural model, thereby obtaining the water body distribution and its boundaries. The adjacency analysis unit is used to analyze adjacency points through distributed boundaries, determine spatial relationships, and obtain a relationship matrix; The random forest calibration unit is used to calibrate the relation matrix by fusing the random forest algorithm if the element value in the relation matrix is ​​greater than zero, and to obtain the bias vector. The distribution boundary adjustment unit is used to adjust the distribution boundary using the deviation vector, determine the calibration deviation, and obtain the calibration result.

7. The system according to claim 1, characterized in that, The iterative optimization module includes: The matching degree index extraction unit is used to process spatial relationship data through deviation calibration, obtain the matching degree index from the fused signal, and obtain the deviation calibration result; Point cloud grouping unit is used to group point cloud data using K-means clustering algorithm if the deviation calibration result shows that the spatial relationship matches the initial fusion signal, so as to obtain a grouped point cloud set; The vector similarity calculation unit is used to calculate the spatial relationship vector based on the grouped point cloud set, determine the vector similarity, and obtain the basis for similarity threshold judgment. The iterative optimization unit is used to adjust the point cloud position through an iterative optimization algorithm to obtain an optimized point cloud framework if the similarity threshold judgment criteria exceed a preset threshold. The feature consistency verification unit is used to fuse the features of the initial fused signal using an optimized point cloud framework, judge the feature consistency, and obtain the consistency verification output. The model draft generation unit is used to update the point cloud construction parameters through consistency verification to obtain the enhanced structural model draft. The surface curve fitting unit is used to fit the surface curves according to the draft of the reinforced structure model using the least squares method to obtain the final reinforced structure model.

8. A method for collecting surveying data for water conservancy projects, characterized in that, include: The distance, angle, and reflection intensity values ​​collected simultaneously are synchronized and weighted to obtain a fused signal; Based on the fused signal, similar features are grouped and data association strength is calculated using a clustering algorithm in a complex environment, and a preliminary point cloud is constructed based on the data association strength. When the data association strength exceeds a preset threshold, a boundary judgment is performed on the preliminary point cloud to obtain the precise coordinate distribution of the engineering boundary; The precise coordinate distribution is input into a neural network to analyze the correspondence between spatial location and reflection intensity, and output the result of the terrain feature continuity judgment. Based on the results of the terrain feature continuity judgment, structural model input data is obtained, the spatial relationship of water body distribution is determined, and the spatial relationship is calibrated for deviation. If the spatial relationship after deviation calibration matches the fused signal, the point cloud is updated through iterative optimization to generate an enhanced structural model. Using the enhanced structural model as a reference, the subsequent acquired signals are processed in real time feedback loop to obtain the final spatial understanding data for water level simulation.

9. An electronic device comprising a memory, a processor, and a computing program stored in the memory and executable on the processor, characterized in that, The processor implements the method of claim 8 when executing the computing program.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of claim 8.

Images