Construction land surveying information management method and system based on internet of things, and medium

Abstract

The application relates to the technical field of information management, in particular to a construction land surveying and mapping information management method and system based on the Internet of Things and a medium. The method comprises the following steps: obtaining construction land surveying and mapping data through the Internet of Things, and performing collaborative sensing fusion to obtain a construction land multi-modal sensing matrix; performing multi-source consistency correction on the construction land multi-modal sensing matrix to obtain a standardized surveying and mapping tensor; obtaining construction land soil data, and analyzing the construction land soil evolution trend based on the construction land soil data; adjusting the semantic weight of the standardized surveying and mapping tensor according to the construction land soil evolution trend and the standardized surveying and mapping tensor to obtain a surveying and mapping enhanced semantic hypergraph; performing land type adaptive sparse coding on the surveying and mapping enhanced semantic hypergraph to obtain real-time surveying and mapping flow data blocks, and uploading the real-time surveying and mapping flow data blocks to a land information management platform for information storage tasks. The application can improve the data retrieval efficiency and accuracy of surveying and mapping data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of information management, and particularly relates to a construction land surveying and mapping information management method and system based on Internet of Things and a medium. BACKGROUND

[0002] The traditional construction land surveying and mapping information management method mainly relies on manual surveying and manual recording, and has low information collection and processing efficiency and large errors and uncertainties. The traditional construction land surveying and mapping management method mainly includes manual surveying, manual recording and traditional geographic information system (GIS) application. Manual surveying needs a large amount of manpower, material resources and time, especially in large-scale surveying and mapping tasks, which often needs to rely on a large number of surveying personnel for on-site operation, not only consumes a large amount of resources, but also has low measurement accuracy. Manual recording and paper archive management also have problems such as information loss, difficulty in sharing and high maintenance cost, which greatly restricts the timeliness and accuracy of data. On the other hand, although the traditional GIS system has good data visualization and management functions in construction land information management, its limitations cannot be ignored. The traditional GIS is usually based on static data, lacks deep combination with real-time dynamic data, and is difficult to realize real-time monitoring and dynamic updating of construction land. In addition, the traditional GIS often has slow data processing speed, insufficient accuracy and poor adaptability to complex terrain and environmental changes when processing large-scale surveying and mapping data. For some key data such as land boundary, geological conditions and land use, the traditional system cannot perform accurate analysis, which affects the decision support function of construction land planning. SUMMARY

[0003] Therefore, it is necessary to provide a construction land surveying and mapping information management method, system and medium based on Internet of Things to solve at least one of the above technical problems.

[0004] To achieve the above-mentioned purpose, a construction land surveying and mapping information management method based on Internet of Things comprises the following steps:

[0005] Step S1: obtaining construction land surveying and mapping data through Internet of Things, and performing collaborative perception fusion on the construction land surveying and mapping data to obtain a construction land multi-modal perception matrix;

[0006] Step S2: performing surveying and mapping data multi-supervision debiasing on the construction land multi-modal perception matrix to obtain a construction land perception tensor, and performing multi-source consistency correction on the construction land perception tensor to obtain a standardized surveying and mapping tensor;

[0007] Step S3: obtaining construction land soil data, and analyzing the construction land soil evolution trend based on the construction land soil data;

[0008] Step S4: According to the evolution trend of the construction land soil and the standardized mapping tensor, the organic matter content causal reasoning is carried out, and the geological organic matter content influence factor is obtained; Based on the geological organic matter content influence factor, the semantic weight of the standardized mapping tensor is adjusted, and the mapping enhanced semantic supergraph is obtained;

[0009] Step S5: The plot type adaptive sparse coding is carried out on the mapping enhanced semantic supergraph, the real-time mapping flow data block is obtained, and is uploaded to the land information management platform for information storage task.

[0010] The present application realizes the real-time collection and multi-modal fusion of construction land surveying and mapping data through Internet of Things technology, improving the breadth and accuracy of data acquisition. Compared with traditional manual surveying and mapping relying on a single measurement method, which leads to limited data coverage and large errors, the present application uses multi-sensor collaborative perception to fuse the data of laser radar, remote sensing images and ground surveying and mapping equipment, constructs a multi-modal perception matrix, thereby enhancing the comprehensiveness and reliability of the surveying and mapping data and reducing the measurement errors that may be caused by a single data source. In addition, through multi-supervised debiasing processing of the surveying and mapping data, systematic biases caused by sensor measurement errors, environmental interference and data noise are eliminated, and the authenticity of the surveying and mapping data is improved. On this basis, multi-source consistency correction technology is used to make different surveying and mapping data sources consistent in spatial coordinates, time synchronization and data scale, solve the information mismatch problem caused by the non-uniformity of data sources in traditional GIS systems, and finally generate standardized surveying and mapping tensors to support subsequent data analysis and modeling. Combined with the analysis of soil data, the modeling and prediction of the soil evolution trend of construction land are realized. Traditional surveying and mapping management methods mainly rely on static soil data, which is difficult to capture the dynamic changes of soil properties, affecting the rational planning of land use. The present application fuses soil moisture, mineral composition, nutrient content and other data to establish a soil evolution trend analysis model, so that the surveying and mapping data are not limited to topographic and geomorphic information, but also take into account geological factors to provide more accurate construction land planning reference. Further, using causal reasoning methods to analyze the influence of soil organic matter content on surveying and mapping characteristics helps to accurately identify key factors affecting geological stability, soil fertility, etc., and adjust the semantic weight of the surveying and mapping data to improve the applicability of the surveying and mapping data. Traditional GIS systems often lack deep mining of complex geographical features when processing geological data. The present application constructs a surveying and mapping enhanced semantic hypergraph, so that the surveying and mapping data can better express the characteristics of land plots in different environments, improving the scientificity of land assessment and management. In terms of data storage and management, the present application uses adaptive sparse coding to optimize land type information, making data storage more efficient and improving the real-time availability of surveying and mapping data. Compared with traditional paper archives or static GIS databases, the present application can update land type information in real time through the Internet of Things and provide more efficient data retrieval and analysis capabilities through the surveying and mapping enhanced semantic hypergraph. In addition, the storage mode of real-time surveying and mapping stream data optimizes the dynamic management capability of surveying and mapping data, enabling the construction land information management platform to continuously obtain the latest surveying and mapping data, providing more accurate decision support for environmental monitoring and other applications, and overcoming the decision delay problem caused by data update lag in traditional GIS systems.

[0011] Optionally, step S1 is specifically:

[0012] Step S11: Obtain construction land surveying and mapping data, and perform data preprocessing on the construction land surveying and mapping data to obtain the construction land surveying and mapping data to be analyzed.

[0013] Step S12: modal decoupling and feature standardization are performed on the construction land surveying and mapping data to be analyzed to obtain standardized surveying and mapping data;

[0014] Step S13: spatiotemporal feature correlation is performed according to the standardized surveying and mapping data to obtain a surveying and mapping spatiotemporal feature map;

[0015] Step S14: multi-modal feature fusion is performed on the surveying and mapping spatiotemporal feature map, and noise surveying and mapping bottlenecks are constrained to obtain a preliminary perception fusion matrix;

[0016] Step S15: information entropy distribution of the preliminary perception fusion matrix is calculated, and low-confidence data regions in the information entropy distribution are subjected to autoregressive completion and regional boundary optimization to obtain a construction land multi-modal perception matrix.

[0017] The present application can effectively clean and standardize the original data, remove measurement errors and noise, and ensure the accuracy of the data in subsequent analysis by obtaining construction land surveying and mapping data and performing data preprocessing. The beneficial effect of this step is that it avoids inaccurate data caused by human error, environmental interference and other factors in traditional surveying and mapping, thereby improving the overall data quality and laying a solid foundation for subsequent processing and analysis. The modal decoupling and feature standardization technology can effectively separate and standardize different modal data (such as laser radar, remote sensing image, etc.) in multi-source data, so that data collected by different sensors can be compared and processed on a unified scale. This technology solves the problem of inconsistency and large standardization difference of different data sources in traditional surveying and mapping methods, improving the consistency and comparability of data. Further, through spatiotemporal feature correlation, a surveying and mapping spatiotemporal feature map is constructed, which can reveal the variation law of data in time and space dimensions, providing deep insight for land planning, environmental monitoring, etc. Through multi-modal feature fusion technology, different types of data sources are combined to enhance the comprehensiveness and depth of surveying and mapping data. At the same time, noise surveying and mapping bottlenecks are constrained to effectively avoid information distortion and noise interference in the data fusion process, ensuring the high quality of the fused data. This process improves the credibility of the data, making subsequent data analysis and decision-making more reliable, and providing more accurate information support for construction land management. By calculating the information entropy distribution of the preliminary perception fusion matrix and performing autoregressive completion and regional boundary optimization on the low-confidence data regions, the present method exhibits significant advantages in data repair and optimization. Especially in the case of missing or abnormal data, the autoregressive completion technology can accurately repair the low-confidence regions, eliminating potential errors caused by incomplete data, thereby enhancing the overall consistency and integrity of the data. In addition, the regional boundary optimization technology further improves the spatial resolution and accuracy of the surveying and mapping data, making the final construction land multi-modal perception matrix more consistent with the actual geographical features, greatly improving the applicability and reliability of the surveying and mapping data.

[0018] Optionally, step S14 is specifically:

[0019] Step S141: Principal component dimension reduction is performed based on the mapping spatiotemporal feature map, feature components with a cumulative contribution rate of greater than or equal to 95% are retained, and a mapping feature subspace is constructed;

[0020] Step S142: Adaptive weight attention multi-modal feature fusion is performed according to the mapping feature subspace, feature contribution degrees of different mapping feature modes in the mapping feature subspace are dynamically adjusted, and a multi-modal weighted fusion feature matrix is obtained;

[0021] Step S143: A low information gain area is detected based on the multi-modal weighted fusion feature matrix, and a Softmax maximum probability less than 0.6 is set to identify a low confidence information area in the multi-modal weighted fusion feature matrix;

[0022] Step S144: High-entropy anomaly detection is performed in combination with the low confidence information area and the low information gain area, a mapping bottleneck mask is obtained by identifying a mapping data noise bottleneck area and performing adaptive multi-scale fusion;

[0023] Step S145: Multi-modal feature bottleneck constraint is performed on the multi-modal weighted fusion feature matrix according to the mapping bottleneck mask, and a preliminary perception fusion matrix is obtained.

[0024] In the present application, the principal component dimension reduction technique retains feature components with a cumulative contribution rate of greater than or equal to 95%, reduces the dimension of data, removes redundant information, simplifies the data structure, and at the same time maintains the key features of the data, thereby laying a more efficient foundation for subsequent analysis. Then, the adaptive weight attention multi-modal feature fusion can dynamically adjust the contribution degrees of different mapping modes, so that the effect of each data type is more accurate, thereby improving the effectiveness and precision of data fusion. Based on the setting of the low information gain area and the Softmax maximum probability threshold, the low confidence information area in the mapping data can be effectively identified, and unreliable data can be effectively screened out, thereby avoiding the interference of low-quality data on the overall result. On this basis, the noise bottleneck area is further identified and located through the high-entropy anomaly detection technology, the accuracy of the data is improved, and the spatial resolution and detail performance of the data are enhanced through adaptive multi-scale fusion. Feature bottleneck constraint is performed in combination with the mapping bottleneck mask, so that the key features in the data are fully retained, and the suppression of noise and redundant parts ensures the quality of the fused data.

[0025] Optionally, step S15 is specifically:

[0026] Step S151: The information entropy distribution of the preliminary perception fusion matrix is calculated, an entropy degree segmentation threshold [0.1, 0.5] is set to divide a high-entropy area and a low-entropy area, and a mapping entropy degree zoning map is obtained.

[0027] Step S152: low-confidence time series interpolation is performed on the low-entropy region in the surveying and mapping entropy degree zoning map, a confidence range of [0.5, 0.8] is set, and a spatial consistency coefficient of [0.1, 0.3] is set to optimize the spatial consistency, and a confidence-enhanced surveying and mapping matrix is obtained;

[0028] Step S153: local anomaly surveying and mapping feature correction is performed on the high-entropy region in the surveying and mapping entropy degree zoning map, and a noise self-adaptive correction matrix is obtained;

[0029] Step S154: region boundary regularization is performed in combination with the confidence-enhanced surveying and mapping matrix and the noise self-adaptive correction matrix, and a boundary-optimized surveying and mapping matrix is obtained;

[0030] Step S155: a consistency evaluation coefficient [0.8, 1.0] is set to perform probability consistency evaluation on the boundary-optimized surveying and mapping matrix, and adaptive confidence re-distribution is performed according to the probability consistency evaluation result, and a construction land multi-modal perception matrix is obtained.

[0031] The present application can effectively distinguish high-entropy and low-entropy regions by information entropy distribution calculation and entropy degree zoning division, so as to identify key differences and uncertainty regions in surveying and mapping data. In the high-entropy region, the data uncertainty is high, and the local anomaly surveying and mapping feature correction can reduce the interference of noise on data quality, further improving the accuracy of the data. In the low-entropy region, through low-confidence time series interpolation combined with spatial consistency optimization, missing data can be effectively filled, and the consistency and reliability of the data can be improved, especially in regions with significant spatial changes. The region boundary regularization in combination with the confidence-enhanced surveying and mapping matrix and the noise self-adaptive correction matrix effectively optimizes the region boundary and reduces the data error caused by unclear boundary. Finally, the probability consistency evaluation of the boundary-optimized surveying and mapping matrix by the consistency evaluation coefficient can automatically adjust the confidence distribution, ensuring high consistency and high accuracy of the data. The combination of these steps improves the quality of the surveying and mapping data, so that the construction land multi-modal perception matrix can more accurately reflect the real plot information.

[0032] Optionally, the surveying and mapping data multi-supervision debiasing in step S2 is specifically:

[0033] Performing hidden variable feature decomposition on the construction land multi-modal perception matrix and performing feature orthogonalization to generate a surveying and mapping hidden variable feature set;

[0034] Performing unsupervised adversarial correction based on the surveying and mapping hidden variable feature set to construct a cross-device distribution adversarial network, and using the cross-device distribution adversarial network to perform maximum mean difference alignment on the construction land multi-modal perception matrix to obtain aligned surveying and mapping data;

[0035] Self-supervised anomaly region detection is performed on the aligned surveying and mapping data, and a preset confidence threshold is set to set the boundary of the anomaly region, to obtain a surveying and mapping anomaly region mask;

[0036] Low-confidence data region error compensation is performed on the aligned surveying and mapping data in combination with the surveying and mapping anomaly region mask, to obtain a construction land perception tensor.

[0037] The present application can effectively extract the core features in the multi-modal perception matrix of construction land by hidden variable feature decomposition and feature orthogonalization, and remove redundant information, thereby improving the expression accuracy of the data. This process helps to capture potential features that are valuable for surveying and mapping data analysis, and enhances the effectiveness of the data in subsequent analysis. The use of unsupervised adversarial correction and cross-device distribution adversarial network for maximum mean difference alignment helps to eliminate the distribution differences between surveying and mapping data from different devices, so that data from different devices are aligned under the same standard, improving the consistency and comparability of the data. Self-supervised anomaly region detection can accurately identify the abnormal regions in the data, and setting the boundary of the abnormal region in combination with the confidence threshold can more accurately define the range of the abnormal region and reduce the error influence. By compensating for the errors of low-confidence data in the abnormal region, the data loss can be effectively filled, the influence caused by data loss or noise can be reduced, and the integrity and accuracy of the surveying and mapping data can be improved. The construction land perception tensor generated finally can more truly and accurately reflect the various features of the construction land, ensuring the efficiency and accuracy of subsequent decision-making and analysis.

[0038] Optionally, the organic matter content causal reasoning in step S4 is specifically:

[0039] Temporal feature decomposition is performed on the construction land soil evolution trend, the evolution features of soil organic matter content are extracted, and time window normalization is performed, to generate a soil evolution feature matrix;

[0040] Evolution feature Granger causality test is performed according to the soil evolution feature matrix, and a preliminary soil organic matter content causal relationship and a preliminary causal weight are set based on the causality test result, to generate a preliminary soil causal graph;

[0041] Perturbation variable analysis is performed on the preliminary soil causal graph, redundant factors are removed, and a causal effect matrix of soil organic matter content is calculated, to construct a causal reasoning model;

[0042] The causal reasoning model is corrected for robustness, and the weight distribution in the corrected causal effect matrix is adjusted through a preset causal entropy weight strategy, to obtain an optimized soil organic matter causal reasoning model;

[0043] The standardized surveying and mapping tensor is inversely deduced based on the optimized soil organic matter causal reasoning model, to quantify the surveying and mapping feature influence factor.

[0044] The application can extract the evolution characteristics of soil organic matter content by time series decomposition and time window normalization, accurately reflect the time series characteristics of soil quality changes, and ensure the timeliness and consistency of the data. This process helps to identify long-term trends and periodic changes in soil organic matter, providing reliable basic data for subsequent analysis. Granger causality test can reveal the causal relationship between soil organic matter content and other variables, providing strong support for understanding the dynamics of soil evolution. According to the results of the causality test, a preliminary soil quality causal diagram is generated, which helps to clarify the influence chain between factors and form a preliminary causal relationship framework. Disturbance variable analysis effectively reduces the complexity of the model by eliminating redundant factors, and accurately quantifies the causal effect of soil organic matter content by calculating the causal effect matrix, improving the accuracy and reliability of the model. The combination of robustness correction and causal entropy weight strategy effectively improves the stability of the model and the ability to cope with abnormal data, and the optimized causal reasoning model can better reflect the change rule of soil organic matter content. Finally, based on the optimized model, the influence factors of the surveying and mapping features can be quantified, thereby providing scientific basis for land resource management.

[0045] Optionally, the semantic weight of the standardized surveying and mapping tensor in step S4 is specifically adjusted as follows:

[0046] A preliminary semantic weight matrix is constructed based on the surveying and mapping feature influence factor and the standardized surveying and mapping tensor;

[0047] The preliminary semantic weight matrix is subjected to soil property correlation weighting analysis to identify the contribution of each surveying and mapping feature in the preliminary semantic weight matrix in reverse deduction, and a surveying and mapping feature deduction contribution is obtained.

[0048] The feature weight in the preliminary semantic weight matrix is adjusted according to the surveying and mapping feature deduction contribution, and an optimized semantic weight matrix is obtained.

[0049] The optimized semantic weight matrix is subjected to gradient descent processing, and the weight distribution is iteratively optimized to obtain a semantic weight distribution optimization matrix.

[0050] The semantic weight distribution optimization matrix is subjected to stability test, and the semantic weight distribution optimization matrix is subjected to adaptive adjustment according to the stability test result to obtain a semantic weight matrix.

[0051] The semantic weight matrix is mapped to the standardized surveying and mapping tensor, the semantic weight of the surveying and mapping feature is adjusted, and a surveying and mapping enhanced semantic supergraph is obtained.

[0052] The application can assign a preliminary semantic weight to each survey feature by constructing a preliminary semantic weight matrix based on survey feature influence factors and standardized survey tensors, providing a clear basis for subsequent optimization. Soil property correlation weighting analysis helps identify the contribution of each survey feature in reverse deduction, accurately assessing the impact of different features on the results. This process ensures the accuracy and reasonableness of the weight matrix, providing a scientific basis for optimized semantic weights. Adjusting the feature weights in the preliminary semantic weight matrix helps better reflect the actual impact of survey data in different environments, ensuring accurate description of soil properties and their changes. Further optimization of weight distribution through gradient descent processing can efficiently improve model performance and prediction accuracy, ensuring convergence and accuracy in the weight adjustment process. Stability testing and adaptive adjustment can enhance the robustness of the model, ensuring stability under different environments or data changes, preventing overfitting or instability. Finally, by mapping the optimized semantic weight matrix to the standardized survey tensor, the semantic weight of the survey feature can be finely adjusted to generate a survey-enhanced semantic hypergraph, improving the accuracy and reliability of land survey data and providing more accurate support for land management decisions.

[0053] Optionally, step S5 is specifically:

[0054] Step S51: clustering survey features based on the survey-enhanced semantic hypergraph to obtain survey feature clustering data;

[0055] Step S52: obtaining plot type expert data, and performing feature matching on the plot type expert data and the survey feature clustering data, assigning plot type labels according to the feature matching to obtain plot type label data;

[0056] Step S53: sparse coding the plot type label data to obtain sparse coding data;

[0057] Step S54: obtaining real-time construction land survey data through the Internet of Things, and performing real-time plot type division and feature coding on the real-time construction land survey data based on the sparse coding data to obtain real-time survey stream data blocks;

[0058] Step S55: uploading the real-time survey stream data blocks to a land information management platform for information storage tasks.

[0059] The application can effectively structure the surveying and mapping data by clustering surveying and mapping features based on the surveying and mapping enhanced semantic hypergraph, accurately reveal the similarity between features, and help identify land features with similar attributes. This process improves the operability and classification accuracy of the data, laying the foundation for land type identification. By matching features with expert land type data, appropriate labels can be assigned to each land type, ensuring the accuracy of the labels and the reliability of the data. The generation of sparse coding data helps to compress redundant data, making the data more concise and efficient, and effectively improving computing and storage efficiency. Real-time acquisition of construction land surveying and mapping data combined with sparse coding for land type division and feature coding enables the surveying and mapping process to reflect land changes in real time, improving the timeliness and accuracy of surveying and mapping data, and ensuring that the land information management platform receives the latest and most accurate data. Finally, uploading real-time surveying and mapping stream data blocks to the platform for information storage can ensure unified management and convenient access of data, thereby improving the overall efficiency of the land information management system and promoting the timeliness and scientificity of decision-making.

[0060] Optionally, the present specification also provides a construction land surveying and mapping information management system based on the Internet of Things, which is used to execute the construction land surveying and mapping information management method based on the Internet of Things as described above. The construction land surveying and mapping information management system based on the Internet of Things comprises:

[0061] A data acquisition module is configured to acquire construction land surveying and mapping data through the Internet of Things, and to perform collaborative perception fusion on the construction land surveying and mapping data to obtain a construction land multi-modal perception matrix.

[0062] A debiasing module is configured to perform surveying and mapping data multi-supervision debiasing on the construction land multi-modal perception matrix to obtain a construction land perception tensor, and to perform multi-source consistency correction on the construction land perception tensor to obtain a standardized surveying and mapping tensor.

[0063] A soil quality evolution trend analysis module is configured to acquire construction land soil quality data, and to analyze the construction land soil quality evolution trend based on the construction land soil quality data.

[0064] A causal reasoning module is configured to perform organic matter content causal reasoning based on the construction land soil quality evolution trend and the standardized surveying and mapping tensor to obtain surveying and mapping feature influence factors, and to adjust the semantic weights of the standardized surveying and mapping tensor based on the surveying and mapping feature influence factors to obtain a surveying and mapping enhanced semantic hypergraph.

[0065] An information storage module is configured to perform land type adaptive sparse coding on the surveying and mapping enhanced semantic hypergraph to obtain real-time surveying and mapping stream data blocks, and to upload the real-time surveying and mapping stream data blocks to a land information management platform for information storage tasks.

[0066] The construction land surveying information management system based on the Internet of Things can realize any one of the construction land surveying information management methods based on the Internet of Things, is used for the medium of joint operation and signal transmission between various modules, and is used for completing the construction land surveying information management method based on the Internet of Things. The internal modules of the system cooperate with each other, so that the management efficiency and accuracy of surveying data management are improved.

[0067] Optionally, the present specification also provides a computer readable storage medium, which stores a computer program, and the computer program is executed to realize the construction land surveying information management method based on the Internet of Things. BRIEF DESCRIPTION OF DRAWINGS

[0068] Other characteristics, objects and advantages of the present application will become more apparent from the following detailed description of non-restrictive embodiments, made with reference to the attached drawings:

[0069] Fig. 1 The step flow diagram of the construction land surveying information management method based on the Internet of Things is shown in the present application;

[0070] Fig. 2 The detailed step flow diagram of step S1 in the present application is shown in the present application;

[0071] Fig. 3 The detailed step flow diagram of step S5 in the present application is shown in the present application;

[0072] The implementation, functional characteristics and advantages of the present application will be further described with reference to the embodiments and the attached drawings. DETAILED DESCRIPTION

[0073] The technical method of the present application will be described clearly and completely below with reference to the attached drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0074] In addition, the drawings are only schematic illustrations of the present application, and are not necessarily drawn to scale. The same reference signs in the drawings represent the same or similar parts, and thus repeated descriptions thereof will be omitted. Some block diagrams shown in the drawings are functional entities, and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0075] It should be understood that, although the terms "first", "second" or the like can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, without departing from the scope of the example embodiments, a first element can be referred to as a second element, and similarly a second element can be referred to as a first element. The term "and / or" as used herein includes any and all combinations of one or more of the associated associated items.

[0076] To achieve the above object, please refer to Figs. 1 to 3 The application provides a construction land surveying information management method based on Internet of Things, which comprises the following steps:

[0077] Step S1: obtaining construction land surveying data through Internet of Things, and performing collaborative perception fusion on the construction land surveying data to obtain a construction land multi-modal perception matrix;

[0078] In this embodiment, the surveying data of multiple sensors is obtained through Internet of Things technology, and the sensors include ground sensors, laser radars, and data photographed by unmanned aerial vehicles. The data generated by each sensor is transmitted to a central processing unit in real time through wireless communication. The data fusion process adopts a weighted average model, and the form of the model is: fusion data matrix = ∑ i = 1 n w i ·X i ; wherein w i represents the weight of the i-th sensor, X i represents the surveying data of the i-th sensor, the weight is determined by the accuracy of the sensor, the value range of w i is [0, 1], and satisfies ∑w i = 1. Finally, the obtained construction land multi-modal perception matrix can be represented as a multi-dimensional matrix M m×n , wherein m is the number of land plots, and n is the dimension of different sensor information of each land plot.

[0079] Step S2: performing multi-supervised deviation removal on the construction land multi-modal perception matrix to obtain a construction land perception tensor, and performing multi-source consistency correction on the construction land perception tensor to obtain a standardized surveying tensor;

[0080] In this embodiment, the perception matrix is subjected to deviation removal processing and multi-source consistency correction. First, a supervised learning algorithm is used to remove the deviation of the construction land multi-modal perception matrix, and a random forest regression model (RF) is adopted. Assuming that the output of the deviation removal model is a matrix M clean, each element of which represents the perception data after bias removal. First, the measurement values of each sensor are standardized, and then the sensor data is corrected using a random forest algorithm to obtain corrected data. Then, multi-source consistency correction is performed, which optimizes the spatio-temporal consistency through a Kalman filter algorithm. The formula of the Kalman filter algorithm is: wherein is the corrected state vector, K k is the Kalman gain, z k is the observation value, H k is the observation matrix. In this process, the construction land perception tensor is corrected, and a more consistent data matrix is obtained. The final output is a standardized mapping tensor, which still has a dimension of m x n, but has removed bias and ensured the consistency of multi-source data.

[0081] Step S3: Obtain construction land soil data, and analyze the construction land soil evolution trend based on the construction land soil data;

[0082] In this embodiment, the construction land soil data is obtained through the Internet of Things, and a regression model is constructed using data such as humidity, pH value, temperature, and organic matter content in the construction land soil data. Time series analysis is performed using an LSTM model (Long Short Term Memory neural network) to predict the evolution trend of the soil. The form of the model is: y t = LSTM(X t-1 , X t-2 ,..., X t-p ); wherein y t is the predicted value at time point t, and X t-p represents the input data at the past p time steps. The input data at each time step includes soil humidity, pH, temperature, etc., and the output y t is the predicted soil change (such as the change trend of humidity). These predicted data can be used to analyze the evolution trend of the soil, and a soil evolution trend matrix M soil trend is generated, which has the form of an m x p matrix, where m is the number of plots and p is the predicted time step length.

[0083] Step S4: Perform organic matter content causal reasoning based on the construction land soil evolution trend and the standardized mapping tensor to obtain geological organic matter content influence factors; adjust the semantic weight of the standardized mapping tensor based on the geological organic matter content influence factors to obtain a mapping enhanced semantic supergraph;

[0084] In this embodiment, the soil evolution trend and the standardized mapping tensor are combined to analyze the influence of organic matter content through a causal reasoning model. Granger causal relationship test and Bayesian network model are used to explore the causal relationship between soil characteristics and organic matter content. The formula of causal reasoning is: y organicf(X soil , X temperature , X humidity ); wherein y organic is the organic matter content, X soil , X temperature , X humidity respectively represent soil types, temperature and humidity and other soil characteristics. The geological organic matter content influence factor matrix is generated using the causal inference result, and the dimension of the matrix is n x 1, wherein n is the number of soil characteristics, and each element represents the influence factor of the corresponding characteristic on the organic matter content. Subsequently, the standardized mapping tensor is weighted and adjusted according to the influence factor. The specific semantic weight adjustment formula is: M adjusted = M standardized ·diag(w); wherein w is the geological organic matter content influence factor, and diag(w) is the weighting of the influence factor converted into a diagonal matrix. Finally, the mapping enhanced semantic supergraph G enhanced is obtained, which represents the complex relationship between plots and contains the correlation between soil characteristics and organic matter content.

[0085] Step S5: The plot type adaptive sparse coding is performed on the mapping enhanced semantic supergraph to obtain real-time mapping stream data blocks, which are uploaded to the land information management platform for information storage tasks.

[0086] In this embodiment, the plot type adaptive sparse coding is performed on the mapping enhanced semantic supergraph, and the K-SVD algorithm is used for sparse coding. Assuming that the formula of the coding process is: X encoded = Sparse(G enhanced , λ); wherein λ is a sparsity parameter that controls the data compression rate after coding. The sparse coded data block X encoded is further divided into real-time mapping stream data blocks and protected by AES encryption to ensure data security. Finally, the mapping data blocks are uploaded to the land information management platform for storage, forming a real-time land data storage system. All mapping data is managed and queried through the information storage platform, ensuring efficient storage and accurate use of land information.

[0087] Alternatively, step S1 is specifically:

[0088] Step S11: Obtain construction land mapping data, and perform data preprocessing on the construction land mapping data to obtain the construction land mapping data to be analyzed;

[0089] In this embodiment, the construction land surveying data of multiple sensors is obtained through an Internet of Things (IoT) platform. These sensors include ground laser radar, image data taken by a drone, temperature and humidity data obtained by ground sensors, and the like. The sensor data is transmitted in real time to a central data processing unit through a wireless network, and the format of the data is a two-dimensional array D raw (size m x n, where m is the number of plots and n is the dimension of different features of each plot). In order to perform data preprocessing, first, the original data is processed for missing values and filtering, and a simple mean interpolation method and Gaussian filtering are used to remove noise. Then, each data source is aligned with a unified timestamp to ensure the consistency of the data in time, thereby obtaining the construction land surveying data to be analyzed. The data matrix D preprocessed after preprocessing has a dimension of m x n, where m is the number of plots and n is the set of all features.

[0090] Step S12: modal decoupling and feature standardization are performed on the construction land surveying data to be analyzed, and standardized surveying data is obtained.

[0091] In this embodiment, modal decoupling is performed on the construction land surveying data D preprocessed to be analyzed. Principal component analysis (PCA) is used to reduce the dimension of the data, and the most representative features in each mode are extracted. In this process, the modal matrix obtained is D modes , where D modes ={D1, D2,..., D p}, each D i corresponds to a data mode, and the dimension is m x n i , where n i is the number of features in this mode. Then, Z-score standardization is performed on each mode, so that the mean of each mode is 0 and the standard deviation is 1. The standardized data matrix is D standardized , which has a dimension of m x n, and the distribution of all features has been adjusted to a unified standard, facilitating subsequent processing.

[0092] Step S13: spatiotemporal feature correlation is performed according to the standardized surveying data, and a surveying spatiotemporal feature map is obtained.

[0093] In this embodiment, spatiotemporal feature correlation is performed based on D standardized . For this purpose, the spatiotemporal graph convolution network (ST-GCN) method is used, and each plot is regarded as a graph node, and the connection between nodes represents the spatial relationship of adjacent plots. In order to represent the time sequence information, we add timestamp information to the feature vector of each node to form a spatiotemporal feature graph G spatial-temporal , and the adjacency matrix A spatial-temporalThe spatial dependency between nodes is described, and the dependency in time is captured by convolution operation. Through the learning of the spatio-temporal convolution network, a spatio-temporal feature map is obtained, which has m nodes and an adjacency matrix of m x m, and each node in the map contains joint features of space and time.

[0094] Step S14: multi-modal feature fusion is performed on the mapping spatio-temporal feature map, and noise mapping bottleneck is constrained to obtain a preliminary perception fusion matrix;

[0095] In this embodiment, multi-modal feature fusion is performed on the mapping spatio-temporal feature map G spatial-temporal . The specific operation is to perform weighted fusion on the features of each mode through a self-attention mechanism (Self-Attention). In this process, a multi-head attention mechanism is used to weight and combine different modal features. Through the learned weight w i and the data D i of each mode, a fused matrix M fusion is obtained, which has a dimension of m x n, where m is the number of plots and n is the feature dimension of each plot. In the fusion process, noise suppression constraints are also used, and L2 regularization and Dropout methods are used to reduce the influence of noise on the final result. Further, a bottleneck constraint is used to suppress noise in the feature map, ensuring that the final perception fusion matrix is more accurate.

[0096] Step S15: calculate the information entropy distribution of the preliminary perception fusion matrix, and perform autoregressive completion and region boundary optimization on the low-confidence data region in the information entropy distribution to obtain a multi-modal perception matrix of construction land.

[0097] In this embodiment, the information entropy distribution of the preliminary perception fusion matrix M fusion is calculated, and autoregressive completion is performed on the low-confidence region in the information entropy distribution. First, the information entropy of each data point in the matrix is calculated, and the formula is: where p(x i ) represents the probability distribution of data point x i , and H(x) is the entropy value of the data point. Through the calculation of the information entropy, a matrix H entropy is obtained, which reflects the confidence of each plot data. Then, autoregressive completion is performed on the low-confidence region. In this step, an ARIMA (Autoregressive Integrated Moving Average) model is used to predict the missing values in the low-confidence region based on historical data. Specifically, the autoregressive model is used to fit the time series data and predict the values of the missing region, and finally a completed matrix M complete is generated. All low-confidence data is repaired and optimized, which is suitable for further data analysis and decision support.

[0098] Optionally, step S14 is specifically:

[0099] Step S141: Principal component dimension reduction is performed based on the mapping spatiotemporal feature map, feature components with a cumulative contribution rate of ≥95% are retained, and a mapping feature subspace is constructed;

[0100] In this embodiment, principal component analysis (PCA) is performed on the mapping spatiotemporal feature map to extract the most representative features and construct a mapping feature subspace. It is assumed that the mapping spatiotemporal feature map G spatial-temporal contains multiple modalities of spatiotemporal features, each modality having a dimension of n i (i = 1, 2,..., p, p is the number of modalities). Through PCA, these features are reduced in dimension, and the goal is to retain feature components with a total cumulative contribution rate of ≥95%. Specifically, the feature matrix G spatial-temporal is subjected to singular value decomposition (SVD): G spatial-temporal = U∑V T ; where U is a feature vector matrix, ∑ is a singular value matrix, and V T is an eigenvalue matrix. By calculating the contribution rate of each principal component, feature components with a cumulative contribution rate of ≥95% are selected to generate a reduced feature matrix G reduced , which has a dimension of m x k, where k is the number of features retained after dimension reduction. The final mapping feature subspace F subspace contains the most representative spatiotemporal features and can be used for subsequent feature fusion operations.

[0101] Step S142: Adaptive weight attention multi-modal feature fusion is performed according to the mapping feature subspace to dynamically adjust the feature contribution of different mapping feature modalities in the mapping feature subspace, and a multi-modal weighted fusion feature matrix is obtained;

[0102] In this embodiment, adaptive weight attention multi-modal feature fusion is performed according to the mapping feature subspace F subspace . To this end, a self-attention mechanism is used to weight and fuse the features of each modality. Specifically, a self-attention mechanism is first applied to each modality's feature vector to calculate the weight a i of each modality. This weight value represents the contribution of the modality feature to the final fusion result. Specifically, a multi-head self-attention mechanism is used to calculate the correlation of each modality feature F i with other modality features: where Q is the query vector, K is the key vector, V is the value vector, and d k is the dimension of the key vector. By calculating the attention weights of each modality, a weighted fusion matrix F fusion, where m is the number of patches, and n is the number of fused features. Finally, the multi-modal weighted fusion feature matrix F fusion The contribution of different modalities can be dynamically adjusted to improve the accuracy of feature fusion.

[0103] Step S143: Detect low information gain regions based on the multi-modal weighted fusion feature matrix, and set Softmax maximum probability < 0.6 to identify low confidence information regions in the multi-modal weighted fusion feature matrix;

[0104] In this embodiment, based on the multi-modal weighted fusion feature matrix F fusion Detect low information gain regions. To achieve this goal, first calculate the information gain of each feature, the formula is: I(X) = H(X)-H(X|Y); where I(X) is the information gain, H(X) is the entropy of feature X, and H(X|Y) is the conditional entropy given feature Y. According to the calculation result, identify the feature region with low information gain. To further identify the low confidence region, use the Softmax function to calculate the maximum probability P max of each feature: P max = max(Softmax(F fusion )); if P max < 0.6, the region is marked as a low confidence region. Finally, the low information gain region and the low confidence region together constitute the detected suspected low information gain region.

[0105] Step S144: Perform high-entropy anomaly detection in combination with low-confidence information regions and low-information gain regions, identify noise bottleneck regions in surveying and mapping data, and perform adaptive multi-scale fusion to obtain a surveying and mapping bottleneck mask;

[0106] In this embodiment, in combination with the low-confidence information region and the low-information gain region, high-entropy anomaly detection is performed to identify noise bottleneck regions in surveying and mapping data. Specifically, first calculate the entropy values of these low-confidence regions and low-information gain regions, and calculate the entropy of each region by formula: High-entropy regions usually indicate anomalies or noise, so they are given special attention. Then, apply a multi-scale fusion method, combine different scale image processing techniques, and use a multi-scale convolutional network to repair noise regions. Specifically, different convolution kernel sizes are used to capture noise features at different scales, thereby achieving efficient noise repair and data optimization. Finally, a surveying and mapping bottleneck mask M mask is obtained, which marks all noise bottleneck regions to help subsequent data repair.

[0107] Step S145: Multi-modal feature bottleneck constraint is performed on the multi-modal weighted fusion feature matrix according to the mapping bottleneck mask to obtain a preliminary perception fusion matrix.

[0108] In this embodiment, the multi-modal feature bottleneck constraint is performed on the multi-modal weighted fusion feature matrix F mask according to the mapping bottleneck mask M fusion . In order to reduce the influence of noise and low-confidence data, a low-confidence region is constrained by a weighted penalty method. Specifically, F final = F fusion × (1-M mask ); where M mask is the mapping bottleneck mask. If a region is a noise region, the value of the corresponding position is 1, otherwise it is 0. Through this constraint, the influence of noise on data can be effectively eliminated, and a more accurate perception fusion result is obtained. Finally, the preliminary perception fusion matrix F final after bottleneck constraint can be used for subsequent analysis and decision support.

[0109] Optionally, step S15 specifically comprises:

[0110] Step S151: The information entropy distribution of the preliminary perception fusion matrix is calculated, an entropy degree segmentation threshold [0.1, 0.5] is set to divide high-entropy regions and low-entropy regions, and a mapping entropy degree zoning map is obtained.

[0111] In this embodiment, the information entropy distribution of the preliminary perception fusion matrix F final is calculated to divide high-entropy regions and low-entropy regions of the mapping data. The information entropy H(X) reflects the uncertainty of the mapping data, and its calculation formula is as follows: where p(x i ) represents the probability distribution of the mapping feature value x i . The local information entropy of each region of F final is calculated, and the entropy degree segmentation threshold range [0.1, 0.5] is set to distinguish high-entropy regions and low-entropy regions. For regions with information entropy values H(X) lower than 0.1, they are marked as low-entropy regions, and regions with information entropy values higher than 0.5 are marked as high-entropy regions, and the remaining regions are not adjusted. Finally, based on the entropy value distribution of the mapping data, a mapping entropy degree zoning map M entropy is constructed, where each grid cell is labeled as a high-entropy (greater noise influence) or low-entropy (higher data consistency) category for subsequent adaptive correction.

[0112] Step S152: Low-confidence time series interpolation is performed on the low-entropy regions in the mapping entropy degree zoning map, a confidence range [0.5, 0.8] is set, and a spatial consistency coefficient [0.1, 0.3] is set to optimize spatial consistency, and a confidence-enhanced mapping matrix is obtained.

[0113] In this embodiment, for the low-entropy region in the mapping entropy partition map M entropy , low-confidence time series interpolation is performed to optimize the spatial consistency of the region. The low-entropy region represents good data quality but may have local missing, so a time series interpolation method is used for repair. Assuming that the confidence distribution of the low-entropy region is C(X), the confidence range is set to [0.5, 0.8], and the mapping data in this range is interpolated. The specific interpolation method uses Bayesian optimal interpolation (BOI), which has the following mathematical form: X interp = X obs + K(X prior -X obs ); where X obs is the observed data, X prior is the prior estimate value, and K is the Kalman gain coefficient, which adjusts the credibility of the interpolated data. In addition, the spatial consistency coefficient range is set to [0.1, 0.3], and the spatial Laplace regularization method is used to optimize the interpolation results to enhance the consistency of the mapping data, and finally generate the confidence-enhanced mapping matrix M conf .

[0114] Step S153: Perform local anomaly mapping feature correction on the high-entropy region in the mapping entropy partition map to obtain a noise-adaptive correction matrix;

[0115] In this embodiment, local anomaly mapping feature correction is performed on the high-entropy region in the mapping entropy partition map M entropy . Since the high-entropy region may have noise, abnormal mapping values or device errors, an adaptive filtering strategy is needed to correct the data. The specific method is to construct an adaptive mean filter (AMF), which has the following update formula: where X i is the mapping value of the high-entropy region, w i is a weight factor related to confidence, and N is the size of the neighborhood window (set to 3x3 or 5x5). When the entropy value of a certain region is high and the mapping value deviates from the neighborhood mean value by more than 2 times the standard deviation, it is considered that the region has an anomaly, and AMF is used for smoothing correction. In addition, to avoid edge effects, a gradient constraint mechanism is used when updating the data, so that the adjusted data will not affect the mapping boundary. Finally, the corrected mapping data form a noise-adaptive correction matrix M denoise .

[0116] Step S154: Perform region boundary regularization on the confidence-enhanced mapping matrix and the noise-adaptive correction matrix to obtain a boundary-optimized mapping matrix;

[0117] In this embodiment, the confidence-enhanced mapping matrix M conf and the noise-adaptive correction matrix M denoise are used to perform boundary regularization to optimize the boundary continuity of the mapping data. Since the confidence-enhanced mapping matrix mainly optimizes low-entropy regions, and the noise-adaptive correction matrix targets high-entropy regions, it is necessary to fuse the two to ensure the smoothness of the data transition. The specific method uses second-order total variation regularization (STV), and the optimization objective is: where X is the boundary-optimized mapping data, and the constraint optimization objective is to reduce data mutations and make the boundary smoother. Subsequently, a boundary gradient constraint is further introduced to ensure that the boundary region is not excessively smoothed: where G boundary is the boundary gradient change. By adjusting the gradient threshold of the boundary region, the integrity of the feature contour can be maintained while eliminating boundary anomalies of the mapping data. Finally, after regularization, the boundary-optimized mapping matrix M boundary is obtained.

[0118] Step S155: Set the consistency evaluation coefficient [0.8, 1.0] to perform probability consistency evaluation on the boundary-optimized mapping matrix, and perform adaptive confidence reassignment according to the probability consistency evaluation result to obtain the construction land multi-modal perception matrix.

[0119] In this embodiment, the probability consistency of the boundary-optimized mapping matrix M boundary is evaluated, and the confidence distribution is adaptively adjusted according to the evaluation result, and finally the construction land multi-modal perception matrix M final is obtained. The consistency evaluation coefficient is set to be in the range of [0.8, 1.0] to measure the consistency of the mapping data. The evaluation method is based on Mahalanobis distance (MD), and the calculation formula is: where μ is the mean of the mapping data, ∑ is the covariance matrix, and X is the data to be evaluated. If the Mahalanobis distance of a certain mapping region is lower than the threshold 0.8, it means that the consistency of this region is higher, and the confidence of this region is increased; if the Mahalanobis distance is higher than 1.0, it means that the data of this region is abnormal, and the confidence is reduced. In order to optimize the final confidence distribution, Dirichlet distribution (Dirichlet Distribution) is used for adaptive confidence adjustment, and the calculation formula is: where C i is the confidence of the mapping data, and α i is the adjustment parameter. Finally, after probability consistency evaluation and adaptive adjustment, the construction land multi-modal perception matrix M final, which is optimized in terms of information consistency, boundary integrity and mapping confidence, can be used for subsequent mapping analysis and planning applications.

[0120] Optionally, the multi-supervised debiasing of the mapping data in step S2 is specifically:

[0121] Performing implicit variable feature decomposition on the multi-modal perception matrix of the construction land and performing feature orthogonalization to generate a mapping implicit variable feature set;

[0122] In this embodiment, the multi-modal perception matrix M final of the construction land is subjected to non-negative matrix factorization (NMF), and the decomposition rank k = 10 is set to extract a low-dimensional implicit variable feature matrix: wherein W represents the implicit variable representation of different mapping modes, and H represents the contribution degree of the implicit variable in different regions. Subsequently, the W is orthogonalized by Gram-Schmidt process: W' = GS(W) to ensure the orthogonality between the implicit variable features and avoid feature redundancy. Finally, a mapping implicit variable feature set M Iv is generated.

[0123] Based on the mapping implicit variable feature set, unsupervised adversarial correction is performed to construct a cross-device distribution adversarial network, and the cross-device distribution adversarial network is used to perform maximum mean difference alignment on the multi-modal perception matrix of the construction land to obtain aligned mapping data;

[0124] In this embodiment, a cross-device adversarial network is constructed for the difference in data distribution of different mapping devices (such as laser radar, satellite remote sensing, and unmanned aerial vehicle mapping), including a generator G(x) and a discriminator D(x). The generator is used to generate mapping data with a uniform distribution, and the discriminator is used to distinguish the data distribution of different devices. The network loss function adopts maximum mean difference (MMD) optimization: wherein φ(x) represents a Gaussian kernel mapping function, and N and M are the number of data samples of different devices. During the training process, the generator is optimized by back propagation to make the distribution of G(x) as close as possible to the distribution of the real mapping data, and finally the aligned mapping data M A is generated.

[0125] Performing self-supervised anomaly region detection on the aligned mapping data and setting an anomaly region boundary in combination with a preset confidence threshold to obtain a mapping anomaly region mask;

[0126] In this embodiment, the anomaly region detection is performed based on a self-supervised Transformer structure. First, a sliding window size w = 5 is used for local feature extraction: F local = CNN(M A , w); then, the local features are input into the Transformer network to calculate an anomaly score S(mi，j ): S(m i，j =Transformer(F) local ); Set the anomaly confidence threshold to 0.9, if S(m i，j If the value is greater than 0.9, then the point is marked as an anomaly. The set of all anomalies constitutes the mapping anomaly area mask M. M .

[0127] By combining the mapping anomaly area mask with the alignment mapping data to perform low-confidence data area error compensation, the construction land perception tensor is obtained.

[0128] In this embodiment, for M A Error compensation is performed on low-confidence regions (i.e., outlier regions) in the model. A Krigin-based approach is used. g Interpolation error compensation strategy to estimate values ​​in low-confidence regions: ∑ k ω k =1; where ω k The data matrix is ​​calculated using Kriging spatial correlation. The compensated data matrix is ​​defined as follows: The obtained construction land sensing tensor It retains the accuracy of high-confidence data while correcting errors in outlier areas.

[0129] Optionally, the causal reasoning of organic matter content described in step S4 specifically includes:

[0130] The soil evolution trend of construction land is decomposed by time series characteristics, the evolution characteristics of soil organic matter content are extracted, and time window normalization is performed to generate a soil evolution characteristic matrix.

[0131] In this embodiment, a time series dataset is constructed for key parameters in the soil evolution trend of construction land, including soil organic carbon (SOC), total nitrogen (TN), and soil moisture content (SWC): S T ={s1, s2, ..., s T}; where s i This represents the soil sample data at the i-th time step. Empirical Mode Decomposition (EMD) is used to analyze S. T Perform temporal feature decomposition to extract evolutionary trends at different scales: Among them, IMF j Let represent the j-th intrinsic mode component, and R be the residual term. Then, principal component analysis (PCA) is performed on the five main eigencomponents (k=5) to reduce dimensionality and obtain the evolution characteristics of soil organic matter content. Next, the locally normalized values ​​are calculated over a sliding time window of 12 months: The normalized data constitute the soil evolution characteristic matrix M Q .

[0132] According to the soil evolution feature matrix, an evolution feature Granger causality test is performed, and a preliminary soil causality graph is generated based on the causality test result, a preliminary causality relationship and a preliminary causality weight of soil organic matter content are set;

[0133] In this embodiment, Granger causality test is used to evaluate the causality between soil variables. The maximum lag order p is set to 3, and a bidirectional test is performed on each feature pair (X, Y) in M Q , the causality strength is calculated by F-statistic, and if the p-value is lower than the significance level (α=0.05), it is determined that there is a Granger causality between variable X and Y, and the causality weight ω(X, Y) is recorded. A preliminary soil causality graph G C =(V, E, W) is constructed, where V represents a set of soil variable nodes, E represents a set of causality edges, and W represents a causality weight matrix.

[0134] A perturbation variable analysis is performed on the preliminary soil causality graph, redundant factors are removed, and a causality effect matrix of soil organic matter content is calculated, thereby constructing a causal reasoning model;

[0135] In this embodiment, a perturbation variable analysis is performed on the causality graph G C . First, the SHAP value (Shapley Additive Explanations) of each variable is calculated by random forest regression, the redundant variables with importance threshold lower than 0.02 are screened, and the corresponding edges E r : G′ C =(V-V r , E-E r , W-W r ) are deleted; where V r is a set of redundant variables, and W r is a weight that is removed. Then, the causality effect matrix E c of soil organic matter content is calculated based on structural equation modeling (SEM), and the coefficients are estimated by least squares: E C =(I-B) -1 Γ; where B is a regression coefficient matrix between variables, and Γ is an influence weight of observation error. A causal reasoning model M c is constructed according to the causality effect matrix, the set of measured variables is set to X′={x1, x2,..., x n}, and the soil organic matter content S O , the measured variable S O =f(X′, E C )+∈; where f(·) is a nonlinear regression function, is the Gaussian noise term. The causal model parameters are optimized using maximum likelihood estimation (MLE) and 5-fold cross-validation to improve the model's generalization ability. Finally, the causal inference model M c can be used to predict the evolution trend of soil organic matter content and provide theoretical support for subsequent steps such as causal effect optimization and mapping feature reverse deduction.

[0136] The causal inference model is robustly corrected, and the weight distribution in the corrected causal effect matrix is adjusted through the preset causal entropy weight strategy to obtain an optimized soil organic matter causal inference model.

[0137] In this embodiment, to enhance the robustness of the causal inference model, Poisson regression (Poisson Regression) is used to correct the causal effect matrix E C for bias: where λ is the Poisson distribution parameter and β is the correction coefficient. Subsequently, the causal entropy weight strategy is used to redistribute the causal weights. Calculate the causal entropy of each variable: H(X i ) = -∑ j p ij logp ij ; then, the entropy weight is assigned to the causal matrix: Finally, the optimized causal inference model M C can more accurately predict the trend of soil organic matter content.

[0138] Based on the optimized soil organic matter causal inference model, the standardized mapping tensor is reverse deduced to quantify the mapping feature influence factor.

[0139] In this embodiment, the optimized causal inference model M C is used to reverse deduce the standardized mapping tensor , where m is the number of mapping area grids and n is the mapping feature dimension (such as terrain slope, elevation, vegetation coverage, etc.). First, based on the optimized causal effect matrix E C , calculate the influence of soil organic matter content S O on mapping features T S through the reverse causal deduction equation: where S O represents the standardized data of soil organic matter content, is the transpose of the optimized causal matrix, indicating the causal influence of soil organic matter content on mapping features. Then, calculate the influence factor matrix I F : where I F represents the normalized relative influence weight of soil organic matter content on each mapping feature. Finally, the influence factor matrix I TThe application can be used to quantify the influence of soil organic matter content changes on topographic features, vegetation index, land surface temperature and other surveying and mapping parameters, and provide accurate data support for soil quality assessment and land use planning.

[0140] Optionally, the semantic weight of the standardized surveying and mapping tensor in step S4 is specifically:

[0141] A preliminary semantic weight matrix is constructed based on the surveying and mapping feature influence factor and the standardized surveying and mapping tensor.

[0142] In this embodiment, the standardized surveying and mapping tensor is wherein m is the number of surveying and mapping area grids, and n is the surveying and mapping feature dimension (such as terrain slope, elevation, vegetation coverage, etc.). The surveying and mapping feature influence factor I F quantifies the influence intensity of each surveying and mapping feature on the soil organic matter content. Through normalization operation, a preliminary semantic weight matrix W O is constructed, and the element is defined as: wherein w ij represents the semantic similarity weight between the surveying and mapping feature i and the surveying and mapping feature j, and the weight range is set between [0, 1].

[0143] The preliminary semantic weight matrix is subjected to soil property correlation weighting analysis to identify the contribution degree of each surveying and mapping feature in the reverse deduction in the preliminary semantic weight matrix, and a surveying and mapping feature deduction contribution degree is obtained.

[0144] In this embodiment, Pearson correlation analysis is used to calculate the correlation between the preliminary semantic weight matrix W O and the soil property variables (such as soil humidity, compactness), and a surveying and mapping feature deduction contribution degree vector C p is defined: wherein, represents the correlation coefficient between the surveying and mapping feature i and the soil organic matter content. If the contribution degree is higher than the threshold value 0.6, it is considered that the feature has an important role in the reverse deduction.

[0145] According to the surveying and mapping feature deduction contribution degree, the feature weight in the preliminary semantic weight matrix is adjusted to obtain an optimized semantic weight matrix.

[0146] In this embodiment, according to the normalized value of C D , the weight of W O is adjusted: W1=W0⊙C P ; wherein, is the element-wise multiplication. In this way, the semantic weight of the high-contribution-degree surveying and mapping feature can be enhanced, and the influence of the low-contribution-degree feature can be weakened, so as to obtain the optimized semantic weight matrix W1.

[0147] The optimized semantic weight matrix is subjected to gradient descent processing, and the weight distribution is iteratively optimized to obtain a semantic weight distribution optimization matrix.

[0148] In this embodiment, the loss function is defined as: wherein, is the updated weight, and λ is a regularization coefficient (such as 0.01). Gradient descent iteration is performed using an Adam optimizer: wherein α is a learning rate (such as 0.05). Finally, a converged semantic weight distribution optimization matrix W2 is obtained.

[0149] The semantic weight distribution optimization matrix is subjected to stability testing, and the semantic weight distribution optimization matrix is adaptively adjusted according to the stability testing result to obtain a semantic weight matrix.

[0150] In this embodiment, the Markov stability analysis method is used to calculate the stability score of the matrix W2 at different time steps: If S t < 0.01, the matrix converges, otherwise the regularization parameter λ is adjusted, and the optimization is continued from the previous step, and finally the semantic weight matrix W3 is obtained.

[0151] The semantic weight matrix is mapped to the standardized surveying and mapping tensor, the semantic weight of the surveying and mapping feature is adjusted, and a surveying and mapping enhanced semantic supergraph is obtained.

[0152] In this embodiment, the standardized surveying and mapping tensor T S is weighted using the semantic weight matrix W3: T' M = T M ⊙W3; a supergraph H is further constructed, wherein the superedge e i is aggregated by surveying and mapping features with high semantic similarity, and a supergraph Laplacian matrix is defined: wherein D H is the superedge degree matrix. Finally, the surveying and mapping enhanced semantic supergraph H is obtained.

[0153] Optionally, step S5 is specifically:

[0154] Step S51: clustering surveying and mapping features based on the surveying and mapping enhanced semantic supergraph to obtain surveying and mapping feature clustering data;

[0155] In this embodiment, the surveying and mapping enhanced semantic supergraph is used to cluster the features of the construction land surveying and mapping data. First, based on the weight relationship in the semantic supergraph, the surveying and mapping features are classified according to the similarity. The K-Means clustering method is used to group the surveying and mapping features in the supergraph to determine the surveying and mapping feature categories of different land plots. In the clustering process, the number of clustering clusters k is set to 5, the attribution category of each feature is iteratively calculated, and finally a surveying and mapping feature clustering data set is formed, and each clustering result represents a surveying and mapping feature set of a land plot.

[0156] Step S52: Obtain plot type expert data and perform feature matching on the plot type expert data and the survey feature clustering data. Assign plot type labels based on the feature matching to obtain plot type label data.

[0157] In this embodiment, the labeled plot type data is obtained from the land and resource database or the artificial survey expert system, including the name of the plot, the purpose (such as farmland, construction land, and forest land), and the survey features. By calculating the similarity between the survey feature clustering data and the expert data, the most matching plot type is selected, and the corresponding plot type label is assigned to the clustered survey data. For example, a certain type of survey feature corresponds to the plot type of "building land", and all data of this type are assigned the label "building land" to form a plot type label data set.

[0158] Step S53: Sparse coding of plot type label data to obtain sparse coding data.

[0159] In this embodiment, the plot type label data is processed using sparse coding. Each plot type is converted to a binary array through the One-Hot Encoding method. For example, the plot type includes "farmland", "building land", and "forest land", and "building land" can be encoded as [0, 1, 0]. This ensures that the data can be effectively matched and calculated in subsequent processing, while reducing the consumption of computing resources and improving data retrieval efficiency.

[0160] Step S54: Obtain real-time construction land survey data through the Internet of Things, and perform real-time plot type division and feature coding on the real-time construction land survey data based on the sparse coding data to obtain real-time survey stream data blocks.

[0161] In this embodiment, real-time construction land survey data is obtained using Internet of Things technology (such as unmanned aerial vehicle surveying, satellite remote sensing, GNSS positioning, etc.). According to the sparse coding data, the survey data obtained in real time is automatically classified to determine its plot type. In the classification process, the similarity calculation method is used to match the survey data to the most similar category and perform feature coding to ensure that the real-time data stream can be quickly associated with the corresponding plot category, and finally generate real-time survey stream data blocks. For example, a survey data point with height and terrain features most similar to "building land" is automatically labeled as "building land" and added to the real-time data stream.

[0162] Step S55: Upload the real-time survey stream data blocks to the land information management platform for information storage tasks.

[0163] In this embodiment, after the real-time mapping flow data block is sorted, it is uploaded to the land information management platform through the data interface. The platform supports structured storage, classifies storage according to land block number, timestamp, land block type and other information, and provides subsequent query and analysis capabilities. For example, the real-time mapping data of a certain land block is uploaded every 10 seconds to ensure timely data updating, providing data support for land planning, management and decision-making.

[0164] Optionally, the present specification also provides a construction land mapping information management system based on Internet of Things, which is used to execute the construction land mapping information management method based on Internet of Things as described above, and the construction land mapping information management system based on Internet of Things comprises:

[0165] A data acquisition module is configured to acquire construction land mapping data through Internet of Things, and to perform collaborative perception fusion on the construction land mapping data to obtain a construction land multi-modal perception matrix.

[0166] A debiasing module is configured to perform mapping data multi-supervision debiasing on the construction land multi-modal perception matrix to obtain a construction land perception tensor, and to perform multi-source consistency correction on the construction land perception tensor to obtain a standardized mapping tensor.

[0167] A soil evolution trend analysis module is configured to acquire construction land soil data, and to analyze a construction land soil evolution trend based on the construction land soil data.

[0168] A causal reasoning module is configured to perform organic matter content causal reasoning based on the construction land soil evolution trend and the standardized mapping tensor to obtain a mapping feature influence factor, and to adjust semantic weights of the standardized mapping tensor based on the mapping feature influence factor to obtain a mapping enhanced semantic hypergraph.

[0169] An information storage module is configured to perform land block type adaptive sparse coding on the mapping enhanced semantic hypergraph to obtain a real-time mapping flow data block, and to upload the real-time mapping flow data block to a land information management platform for an information storage task.

[0170] Optionally, the present specification also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed to implement the construction land mapping information management method based on Internet of Things as described above.

[0171] Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, and the scope of the present application is defined by the appended claims rather than the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the application file are intended to be included in the present application.

[0172] The foregoing is considered as illustrative only of the principles of the application. Numerous modifications and changes will readily occur to those skilled in the art, and it is intended to embrace all such modifications and changes that fall within the scope of the application. Accordingly, the application is not to be restricted in scope to the specific embodiments disclosed herein but is to be accorded the full scope that the principles and novel features request appropriately granted.

Claims

1. A method for managing construction land surveying and mapping information based on the Internet of Things, characterized in that, Includes the following steps: Step S1: Acquire construction land surveying data through the Internet of Things, and perform collaborative sensing and fusion of construction land surveying data to obtain a construction land multimodal sensing matrix; Step S2: Perform multi-supervised debiasing on the construction land multimodal sensing matrix to obtain the construction land sensing tensor, and perform multi-source consistency correction on the construction land sensing tensor to obtain the standardized mapping tensor; Step S3: Obtain soil data for construction land and analyze the soil evolution trend based on the soil data; Step S4: Based on the soil evolution trend of construction land and the standardized mapping tensor, perform causal inference on organic matter content to obtain the mapping feature influence factor; adjust the semantic weight of the standardized mapping tensor based on the mapping feature influence factor to obtain the mapping-enhanced semantic hypergraph; the causal inference on organic matter content specifically involves: The soil evolution trend of construction land is decomposed by time series characteristics, the evolution characteristics of soil organic matter content are extracted, and time window normalization is performed to generate a soil evolution characteristic matrix. Based on the soil evolution feature matrix, Granger causality test of evolution features is performed. Based on the causality test results, preliminary causal relationship and preliminary causal weight of soil organic matter content are set, thereby generating a preliminary soil causality map. The preliminary soil causal diagram was subjected to perturbation variable analysis to remove redundant factors, and the causal effect matrix of soil organic matter content was calculated to construct a causal inference model. The causal reasoning model is robustly corrected, and the weight distribution in the corrected causal effect matrix is ​​adjusted by a preset causal entropy weight strategy to obtain an optimized causal reasoning model for soil organic matter. Based on the optimized causal reasoning model of soil organic matter, the standardized mapping tensor is inversely deduced to quantify the influencing factors of mapping characteristics; The adjustment of the semantic weights of the standardized mapping tensor specifically refers to: A preliminary semantic weight matrix is ​​constructed based on the influencing factors of surveying and mapping features and the standardized surveying and mapping tensor. A soil property correlation weighted analysis was performed on the preliminary semantic weight matrix to identify the contribution of each mapping feature in the reverse inference and obtain the mapping feature inference contribution. The feature weights in the initial semantic weight matrix are adjusted based on the contribution of the mapping features to obtain the optimized semantic weight matrix. Gradient descent is applied to the optimized semantic weight matrix to iteratively optimize the weight distribution and obtain the optimized semantic weight distribution matrix. The semantic weight distribution optimization matrix is ​​subjected to stability test, and the semantic weight distribution optimization matrix is ​​adaptively adjusted according to the stability test results to obtain the semantic weight matrix; By mapping the semantic weight matrix to the standardized mapping tensor and adjusting the semantic weights of the mapping features, a mapping-enhanced semantic hypergraph is obtained. Step S5: Perform land parcel type adaptive sparse coding on the mapping augmented semantic hypermap to obtain real-time mapping stream data blocks, and upload them to the land information management platform for information storage.

2. The method for managing construction land surveying and mapping information based on the Internet of Things according to claim 1, characterized in that, Step S1 is as follows: Step S11: Obtain construction land survey data and perform data preprocessing on the construction land survey data to obtain the construction land survey data to be analyzed; Step S12: Perform modal decoupling and feature standardization on the construction land survey data to be analyzed to obtain standardized survey data; Step S13: Perform spatiotemporal feature correlation based on standardized surveying and mapping data to obtain a surveying and mapping spatiotemporal feature map; Step S14: Perform multimodal feature fusion on the spatiotemporal feature map and constrain the bottleneck of noisy mapping to obtain a preliminary perception fusion matrix; Step S15: Calculate the information entropy distribution of the preliminary perception fusion matrix, and perform autoregressive completion and regional boundary optimization on the low-confidence data areas in the information entropy distribution to obtain the multimodal perception matrix of construction land.

3. The method for managing construction land surveying and mapping information based on the Internet of Things according to claim 2, characterized in that, Step S14 is as follows: Step S141: Perform principal component dimensionality reduction based on the spatiotemporal feature map, retain feature components with a cumulative contribution rate of ≥95%, and construct the mapping feature subspace; Step S142: Perform adaptive weighted attention multimodal feature fusion based on the mapping feature subspace, dynamically adjust the feature contribution of different mapping feature modes in the mapping feature subspace, and obtain the multimodal weighted fusion feature matrix; Step S143: Detect low information gain regions based on the multimodal weighted fusion feature matrix, and set the maximum probability of Softmax to <0.6 to identify low confidence information regions in the multimodal weighted fusion feature matrix; Step S144: Combine low-confidence information regions and low-information-gain regions to perform high-entropy anomaly detection, identify the noise bottleneck region of the mapping data, and perform adaptive multi-scale fusion to obtain the mapping bottleneck mask. Step S145: Apply multimodal feature bottleneck constraints to the multimodal weighted fusion feature matrix based on the mapping bottleneck mask to obtain the preliminary perception fusion matrix.

4. The method for managing construction land surveying and mapping information based on the Internet of Things according to claim 2, characterized in that, Step S15 is as follows: Step S151: Calculate the information entropy distribution of the preliminary perception fusion matrix, set the entropy segmentation threshold [0.1, 0.5] to divide the high-entropy region into the low-entropy region, and obtain the mapping entropy partition map; Step S152: Perform low-confidence time series interpolation on the low-entropy region in the mapping entropy partition map, set the confidence range to [0.5, 0.8], and set the spatial consistency coefficient to [0.1, 0.3] to optimize spatial consistency, and obtain the confidence-enhanced mapping matrix; Step S153: Correct the local anomaly mapping features of the high-entropy region in the mapping entropy partition map to obtain the noise adaptive correction matrix; Step S154: Combine the confidence enhancement mapping matrix and the noise adaptive correction matrix to perform region boundary regularization, and obtain the boundary optimization mapping matrix; Step S155: Set the consistency evaluation coefficient [0.8, 1.0] to perform probabilistic consistency evaluation on the boundary optimization mapping matrix, and perform adaptive confidence redistribution based on the probabilistic consistency evaluation results to obtain the multimodal perception matrix of construction land.

5. The method for managing construction land surveying and mapping information based on the Internet of Things according to claim 1, characterized in that, The multi-supervised bias correction of the surveying data mentioned in step S2 specifically refers to: Latent variable feature decomposition is performed on the multimodal perception matrix of construction land, and feature orthogonalization is performed to generate a set of latent variable features for surveying and mapping. Unsupervised adversarial correction is performed based on the latent variable feature set of surveying and mapping to construct a cross-device distributed adversarial network. The cross-device distributed adversarial network is then used to perform maximum mean difference alignment on the multimodal perception matrix of construction land to obtain aligned surveying and mapping data. Self-supervised anomaly detection is performed on the aligned mapping data, and the anomaly region boundary is set in combination with the preset confidence threshold to obtain the mapping anomaly region mask; By combining the mapping anomaly area mask with the alignment mapping data to perform low-confidence data area error compensation, the construction land perception tensor is obtained.

6. The method for managing construction land surveying and mapping information based on the Internet of Things according to claim 1, characterized in that, Step S5 is as follows: Step S51: Perform mapping feature clustering based on the mapping augmented semantic hypergraph to obtain mapping feature clustering data; Step S52: Obtain land parcel type expert data, and perform feature matching on the land parcel type expert data and survey feature clustering data. Assign land parcel type labels based on feature matching to obtain land parcel type label data. Step S53: Perform sparse coding on the land parcel type label data to obtain sparse coded data; Step S54: Obtain real-time construction land surveying data through the Internet of Things, and perform real-time land parcel type classification and feature encoding on the real-time construction land surveying data based on sparse coded data to obtain real-time surveying stream data blocks; Step S55: Upload the real-time mapping stream data block to the land information management platform for information storage.

7. A construction land surveying and mapping information management system based on the Internet of Things, characterized in that, For executing the Internet of Things-based construction land surveying and mapping information management method as described in claim 1, the Internet of Things-based construction land surveying and mapping information management system includes: The data acquisition module is used to acquire construction land surveying data through the Internet of Things, and to perform collaborative sensing and fusion of the construction land surveying data to obtain a construction land multimodal sensing matrix; The bias correction module is used to perform multi-supervised bias correction on the multimodal sensing matrix of construction land to obtain the construction land sensing tensor, and to perform multi-source consistency correction on the construction land sensing tensor to obtain the standardized surveying tensor. The soil evolution trend analysis module is used to acquire soil data of construction land and analyze the soil evolution trend of construction land based on the soil data. The causal reasoning module is used to perform causal reasoning on organic matter content based on the soil evolution trend of construction land and standardized surveying tensor to obtain surveying feature influence factors; and to adjust the semantic weights of standardized surveying tensor based on surveying feature influence factors to obtain a surveying-enhanced semantic hypergraph. The information storage module is used to perform adaptive sparse coding of land parcel types on the surveying and mapping augmented semantic hypermap to obtain real-time surveying and mapping stream data blocks, and upload them to the land information management platform for information storage tasks.

8. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed, implements the Internet of Things-based construction land surveying and mapping information management method as described in any one of claims 1-6.

Images