Natural resource analysis method and system combined with multi-source data

Abstract

The application provides a natural resource analysis method and system combining multi-source data, first acquires a multi-source data set of remote sensing images, geographic monitoring and environmental monitoring and the like of a target area, then carries out spatio-temporal feature extraction processing, obtains a resource space distribution feature set and a time change feature set, accurately depicts the features of the resources in different spatio-temporal dimensions, then carries out dynamic correlation processing on the resource space distribution feature set and the resource time change feature set based on a preset dynamic correlation rule set, generates a resource dynamic correlation relationship set, reveals the interaction and evolution law between the resources, generates a resource state evaluation strategy according to the resource dynamic correlation relationship set, and determines a resource management optimization direction, finally, feeds back the resource management optimization direction to a resource management system, triggers resource scheduling operation, realizes scientific, accurate and dynamic management of natural resources, and improves resource management efficiency and sustainability.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a natural resource analysis method and system combining multi-source data. Background Art

[0002] In the field of natural resource management, with the continuous expansion of human activities and the increasing intensity of resource development and utilization, accurate analysis and scientific management of natural resources have become increasingly important. However, existing natural resource analysis technologies have many limitations and cannot meet the complex and changing needs of resource management.

[0003] Currently, most natural resource analyses rely on static methods, simply integrating and overlaying data from multiple sources while ignoring the temporal and spatial connections and dynamic changes between data. This static approach fails to capture the interactions and evolution of resources across different temporal and spatial scales, leading to significant deviations between analytical results and actual conditions, making it difficult to provide a scientific and accurate basis for resource management decisions.

[0004] Furthermore, existing natural resource analysis methods lack effective assessments of resource status and targeted management optimization strategies. Even if sufficient resource information is available, it is difficult to formulate appropriate resource management plans based on this information, making it impossible to dynamically schedule and optimize resource allocation. In resource management systems, the lack of effective linkage with analysis results prevents timely adjustments to management strategies based on changes in resource status, leading to inefficient resource management, frequent resource waste, and environmental damage. Summary of the Invention

[0005] In view of the above-mentioned problems, in combination with the first aspect of the present invention, an embodiment of the present invention provides a natural resource analysis method combining multi-source data, the method comprising:

[0006] Acquire a multi-source data set formed by multiple data sources of a target area, wherein the multi-source data set includes a remote sensing image data set, a geographic monitoring data set, and an environmental monitoring data set;

[0007] Performing spatiotemporal feature extraction processing on the multi-source data set to obtain a resource spatial distribution feature set and a resource temporal change feature set of the target area;

[0008] Based on a preset dynamic association rule set, dynamically associate the resource spatial distribution feature set and the resource time change feature set to generate a resource dynamic association relationship set;

[0009] Generating a resource status evaluation strategy according to the resource dynamic association relationship set, and determining a resource management optimization direction based on the resource status evaluation strategy;

[0010] The resource management optimization direction is fed back to the resource management system to trigger resource scheduling operations.

[0011] On the other hand, an embodiment of the present invention also provides a natural resource analysis system combining multi-source data, including a processor and a machine-readable storage medium, wherein the machine-readable storage medium is connected to the processor, the machine-readable storage medium is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the machine-readable storage medium to implement the above method.

[0012] Based on the above aspects, the embodiment of the present invention integrates multi-source data sets such as remote sensing images, geographic monitoring and environmental monitoring of the target area, and then performs spatiotemporal feature extraction and processing on the multi-source data sets, accurately depicting the spatial distribution characteristics and temporal change characteristics of the resources, and dynamically associates the spatiotemporal features based on a preset dynamic association rule set to generate a resource dynamic association relationship set. The resource dynamic association relationship set can truly reflect the interaction and evolution laws of resources in different spatiotemporal dimensions. The resource status assessment strategy generated according to the resource dynamic association relationship set can accurately judge the current status of resources and determine targeted resource management optimization directions, making resource management more scientific and forward-looking, and feeding back the resource management optimization direction to the resource management system to trigger resource scheduling operations, thereby realizing closed-loop control from data analysis to resource management decision-making, and significantly improving the efficiency, accuracy and intelligence level of resource management. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a schematic diagram of the execution flow of the natural resource analysis method combining multi-source data provided by an embodiment of the present invention.

[0014] Figure 2 Schematic diagram of exemplary hardware and software components of a natural resource analysis system combining multi-source data provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0015] The present invention will be described in detail below with reference to the accompanying drawings. Figure 1 This is a flow chart of a natural resource analysis method combining multi-source data provided by an embodiment of the present invention. The natural resource analysis method combining multi-source data is introduced in detail below.

[0016] Step S110: obtaining a multi-source data set formed by multiple data sources of the target area, wherein the multi-source data set includes a remote sensing image data set, a geographic monitoring data set, and an environmental monitoring data set.

[0017] To accurately and comprehensively understand the natural resource status of the target area, it is necessary to obtain relevant data from multiple different data sources. Remote sensing image data sets can be obtained from satellite remote sensing platforms. For example, if the satellite platform has M satellites, each satellite collects image data of N different bands for the target area. Let the image data of the jth band collected by the i-th satellite be R_ij, where i ranges from 1 to M and j ranges from 1 to N. These image data have different spatial resolutions and data formats, and together constitute the remote sensing image data set R = {R_ij|i = 1, 2, ..., M; = 1, 2, ..., N}. Image data of different bands can reflect different surface characteristics of the target area. For example, the visible light band can show the color and texture of the surface, while the infrared band can help detect the health of vegetation.

[0018] The acquisition of a geographic monitoring data set relies on P geographic monitoring stations distributed throughout the target area. Each monitoring station collects various types of geographic information, including terrain elevation and soil type information. Let the terrain elevation data collected by the kth monitoring station be G_1k, and the soil type data be G_2k, where k ranges from 1 to P. The geographic monitoring data set G can then be expressed as G = {G_1k, G_2k | k = 1, 2, ..., P}. Terrain elevation data reflects the topography of the target area, while soil type data helps understand soil properties and distribution.

[0019] Environmental monitoring data sets are collected by Q environmental monitoring stations distributed across the target area. Each monitoring station monitors multiple environmental indicators in real time, and these indicators change over time. Let the mth environmental indicator data collected by the lth monitoring station at the tth time be E_ltm, where l ranges from 1 to Q, t represents a different time point, and m ranges from 1 to the number of environmental indicators. The environmental monitoring data set E can be expressed as E={E_ltm|l=1, 2, ..., Q; t is the time point; m is the type of environmental indicator}. Environmental indicators may include air quality indicators (such as pollutant concentrations) and water quality indicators (such as pH and dissolved oxygen). These data can reflect the changes in the environmental quality and ecological status of the target area over time.

[0020] Step S120: performing spatiotemporal feature extraction processing on the multi-source data set to obtain a resource spatial distribution feature set and a resource temporal change feature set of the target area.

[0021] Since the data in the multi-source data set have different formats, resolutions, and dimensions, in order to extract useful spatiotemporal features from them, these multi-source data need to be preprocessed and analyzed. Specifically, the following sub-steps are included:

[0022] Step S121: performing spatial resolution alignment processing and unit uniform conversion processing on the remote sensing image data set to generate a standardized remote sensing image data set.

[0023] In the acquired remote sensing image dataset R, different image data R_ij may not only have different spatial resolutions, but their raw data may also contain sensor-specific radiometric values, such as DN. To ensure the accuracy of subsequent processing and analysis, radiometric calibration and atmospheric correction are required before spatial resolution alignment and unit conversion.

[0024] Radiometric calibration is the process of converting the raw digital quantization values ​​(DN values) recorded by a sensor into physically meaningful radiance values. For each image data point R_ij, the DN value is converted to a radiance value by looking up the sensor's calibration parameters. Assuming that the calibration parameters for the image data of the jth band collected by the i-th satellite are a_ij and b_ij, the radiometrically calibrated image data point R_cal_ij can be calculated as follows: first, multiply each DN value in R_ij by a_ij, then add b_ij to convert the DN value to a radiance value.

[0025] Atmospheric correction is the process of eliminating the atmospheric influence on the radiative transmission of remote sensing images. Because gas molecules and aerosols in the atmosphere scatter and absorb solar radiation, the radiance values ​​received by the sensor may not truly reflect the surface's reflectance. Using an appropriate atmospheric correction model, such as the 6S model or the FLAASH model, the radiometrically calibrated image data (R_cal_ij) is processed and calibrated to a surface reflectance or radiance standard, resulting in the atmospherically corrected image data (R_atm_ij).

[0026] After completing radiometric calibration and atmospheric correction, spatial resolution alignment is performed. A reference resolution is selected as a unified standard, assuming it is R_ref. For each atmospherically corrected image data R_atm_ij, a bilinear interpolation algorithm is used to adjust its resolution to be consistent with R_ref, resulting in the adjusted image data R'_ij.

[0027] While performing spatial resolution alignment, the image data units also need to be uniformly converted. All image data units are converted to a unified unit, U, such as radiance units. Through a series of conversion operations, the units of each adjusted image data R'_ij are converted to U, resulting in standardized image data R''_ij. Ultimately, these standardized image data constitute the standardized remote sensing image data set R_std = {R''_ij | i = 1, 2, ..., M; j = 1, 2, ..., N}.

[0028] Step S122: calling a pre-trained semantic segmentation model to perform surface cover type recognition processing on the standardized remote sensing image data set to obtain a surface cover distribution feature set of the target area.

[0029] The pre-trained semantic segmentation model is based on deep learning and typically consists of an input layer, multiple convolutional layers, a pooling layer, an upsampling layer, and an output layer. To fully exploit the correlation of multispectral information, the model's input should be a multi-band fused image. The image data from different bands at the same location in the standardized remote sensing image dataset R_std is fused to form a multi-band fused image. Assuming that each location has N bands of image data, these N bands of data are used as input channels to construct an input tensor that is fed into the input layer of the semantic segmentation model.

[0030] In the convolutional layer, the model extracts image feature information through a series of convolution kernels. Each kernel detects different characteristic patterns in the image, such as edges and textures. The pooling layer reduces the dimensionality of the feature map, reducing computational effort while retaining important feature information. The upsampling layer restores the feature map to the same size as the input image, facilitating pixel-level classification. The output layer outputs the land cover type label for each pixel.

[0031] Assume there are S types of land cover, denoted by T_1, T_2, …, T_S. A semantic segmentation model processes the standardized multi-band fused remote sensing image data, assigning a land cover type label to each pixel. Combining these labels yields a land cover type map corresponding to the image. After processing all the multi-band fused image data, the resulting land cover type maps are merged to form a set of land cover distribution features for the target area: C = {C_ij | i = 1, 2, …, M; j = 1, 2, …, N}, where C_ij represents the land cover type map corresponding to the multi-band fused image acquired by the i-th satellite.

[0032] Step S123: performing spatial reference coordinate system conversion and unit unification processing on the geographic monitoring data set to generate a terrain elevation distribution feature set and a soil type distribution feature set consistent with the resolution of the standardized remote sensing image.

[0033] The data in the geographic monitoring dataset G may use different spatial reference coordinate systems. To match and fuse them with the standardized remote sensing image dataset R_std, the geographic monitoring data needs to be converted to a spatial reference coordinate system. First, a unified spatial reference coordinate system is determined, assuming it is CS_ref. For each terrain elevation data G_1k and soil type data G_2k in the geographic monitoring dataset G, a coordinate conversion algorithm is used to convert them from their original spatial reference coordinate system to CS_ref, resulting in the converted terrain elevation data G'_1k and soil type data G'_2k.

[0034] Terrain elevation data is a continuous variable and requires unit unification, converting its units to the unit of length, L. Soil type data, on the other hand, is a categorical variable and inherently does not require unit conversion. After completing the spatial reference coordinate system conversion and unit unification of the terrain elevation data, the resolution of both the terrain elevation data and the soil type data must be adjusted to match the resolution of the standardized remote sensing image dataset, R_std. This resolution adjustment can be achieved through interpolation and resampling.

[0035] After processing, we obtain the terrain elevation distribution feature set H = {H_k|k = 1, 2, …, P} and the soil type distribution feature set S_type = {S_k|k = 1, 2, …, P}, where H_k represents the terrain elevation distribution feature corresponding to the kth monitoring station, and S_k represents the soil type distribution feature corresponding to the kth monitoring station. To prevent the model from mistakenly treating categorical labels as continuous values ​​when subsequently merging spatial features, the data in the soil type distribution feature set S_type must be one-hot encoded to clearly mark them as categorical variables.

[0036] Step S124: performing time series alignment processing and granularity unification processing on the environmental monitoring data set to generate a standardized environmental monitoring data set.

[0037] The data in the environmental monitoring data set E varies over time, and different monitoring stations may have different sampling intervals and data recording methods. To facilitate subsequent time series analysis, the environmental monitoring data must be time-series aligned. First, a unified time base and time interval are determined. Assume that the time base is t_0 and the time interval is Δt. For each monitoring station's environmental indicator data, E_ltm, it is sorted chronologically and interpolated and resampled based on the time base and time interval to align all data in time.

[0038] Furthermore, the granularity (i.e., the level of detail) of environmental monitoring data may vary, necessitating uniform granularity. Appropriate interpolation methods are selected based on the characteristics of the environmental indicators. For example, for air quality indicators, the maximum retention method is used to interpolate, as it is important to focus on peak values ​​associated with sudden pollution events. For water quality data, a sliding window averaging method is used to more smoothly reflect changes in water quality.

[0039] Suppose we use a sliding window averaging method to uniformly process the granularity of environmental indicator data, with a window size of W. For each time point t, the average value of the data within the window of size W centered at that time point is calculated as the new value at that time point. In this way, the originally hourly data is uniformly recorded on a daily basis, completing the granularity uniformity process.

[0040] After time series alignment and granularity unification, the standardized environmental monitoring data set E_std={E'_ltm|l=1, 2,…, Q; t is a unified time point; m is the type of environmental indicator} is obtained.

[0041] Step S125: calling a pre-trained time series analysis model to perform periodic change pattern extraction processing on the standardized environmental monitoring data set to obtain a resource time change feature set of the target area.

[0042] Pre-trained time series analysis models can employ recurrent neural networks (RNNs) and their variants, such as long short-term memory (LSTM) networks or gated recurrent units (GRUs). These models are capable of processing temporally ordered data and capturing cyclical patterns within the data. The structure of a time series analysis model typically consists of an input layer, hidden layers, and an output layer. The input layer receives data E'_ltm from the standardized environmental monitoring dataset E_std. Neurons in the hidden layer memorize past information through recurrent connections and update their own states based on current input and past states. The output layer outputs the cyclical variation characteristics of each environmental indicator.

[0043] When extracting periodic patterns, it's necessary to determine the adaptive time window length. This window length is determined by combining domain knowledge with an automatic period detection algorithm. For example, for climate-related environmental indicators, given that the monsoon cycle typically lasts six months, the window length can be initially set to a time step corresponding to six months. Furthermore, an automatic period detection algorithm, such as the Fourier transform (FFT), is used to analyze the standardized environmental monitoring data set E_std. The main periodic components in the data are identified based on the spectrum, allowing for further adjustment of the window length.

[0044] For each environmental indicator data point E'_ltm in the standardized environmental monitoring data set E_std, the time series analysis model processes it according to the determined adaptive time window length, extracting the cyclical variation pattern of the indicator. These patterns can be described using characteristics such as cycle length, amplitude, and phase. By combining the cyclical variation pattern characteristics of all environmental indicators, we obtain the resource temporal variation feature set T = {T_lm | l = 1, 2, …, Q; m is the type of environmental indicator} for the target area, where T_lm represents the cyclical variation pattern characteristics of the mth environmental indicator collected by the lth monitoring station.

[0045] Step S126: merging the surface cover distribution feature set, the terrain elevation distribution feature set, and the soil type distribution feature set into the resource space distribution feature set.

[0046] The land cover distribution feature set C, the terrain elevation distribution feature set H, and the soil type distribution feature set S_type are merged to form the resource spatial distribution feature set D. The merging process can be done in a splicing manner, that is, the corresponding elements in the three feature sets are combined. For example, for the land cover distribution feature C_k, the terrain elevation distribution feature H_k, and the soil type distribution feature S_k corresponding to the k-th monitoring station, they are spliced ​​together to obtain the resource spatial distribution feature D_k corresponding to the monitoring station. By combining the resource spatial distribution features of all monitoring stations, the resource spatial distribution feature set D = {D_k|k = 1, 2, ..., P} is obtained.

[0047] Step S130: Based on a preset dynamic association rule set, dynamically associate the resource spatial distribution feature set and the resource temporal change feature set to generate a resource dynamic association relationship set.

[0048] In order to gain a deeper understanding of the inherent connections and changing patterns of natural resources in the target area, it is necessary to dynamically associate the resource spatial distribution feature set D and the resource temporal change feature set T according to a preset dynamic association rule set. This specifically includes the following sub-steps:

[0049] Step S131: obtaining a first association rule subset in the dynamic association rule set, where the first association rule subset is used to characterize a spatial association constraint condition between land cover types and terrain elevations.

[0050] The dynamic association rule set contains multiple different association rule subsets, which are used to describe the associations between different types of features. The first association rule subset R1 describes the spatial association constraints between land cover types and terrain elevation. These constraints can be based on geographical knowledge and experience. For example, certain land cover types (such as forests) are typically distributed within a specific terrain elevation range. The first association rule subset R1 is extracted from the dynamic association rule set.

[0051] Step S132: performing resolution consistency check on the surface cover distribution feature set and the terrain elevation distribution feature set according to the first association rule subset, performing spatial matching processing in a unified grid coordinate system, and generating a first association relationship subset.

[0052] Before performing spatial matching, it is necessary to perform a resolution consistency check on the land cover distribution feature set C and the terrain elevation distribution feature set H. Ensure that the resolutions of these two feature sets are the same so that accurate matching can be performed in a unified grid coordinate system. If the resolutions are inconsistent, interpolation or resampling can be used to adjust them.

[0053] In a unified grid coordinate system, spatial matching is performed on the land cover distribution feature set C and the terrain elevation distribution feature set H according to the first association rule subset R1. For each grid cell, the land cover type and terrain elevation within the cell are checked to see if they meet the constraints in the first association rule subset R1. If so, the association between the land cover type and terrain elevation for that cell is recorded to form the first association relationship subset A1.

[0054] Step S133: obtaining a second association rule subset in the dynamic association rule set, where the second association rule subset is used to characterize spatial compatibility constraints between soil types and land cover types.

[0055] A second association rule subset R2 is extracted from the dynamic association rule set. This second association rule subset is used to describe the spatial compatibility constraints between soil types and land cover types. These constraints reflect the suitability of different soil types for land cover types. For example, some soil types are more suitable for growing specific vegetation.

[0056] Step S134: performing spatial overlay analysis on the unified soil type distribution feature set and the land cover distribution feature set according to the second association rule subset to generate a second association relationship subset.

[0057] Perform spatial overlay analysis on the unified soil type distribution feature set S_type and the land cover distribution feature set C. Overlay these two feature sets spatially. For each spatial location, check whether the soil type and land cover type at that location satisfy the spatial compatibility constraints in the second association rule subset R2. If so, record the association between the soil type and land cover type at that location to form the second association relationship subset A2.

[0058] Step S135: obtaining a third association rule subset in the dynamic association rule set, wherein the third association rule subset is used to characterize a temporal response constraint condition between environmental monitoring indicators and resource time variation characteristics.

[0059] A third association rule subset, R3, is extracted from the dynamic association rule set. This subset describes the temporal response constraints between environmental monitoring indicators and resource temporal variation characteristics. These constraints reflect how changes in environmental monitoring indicators affect the temporal variation characteristics of resources. For example, changes in air quality may affect the growth cycle of vegetation.

[0060] Step S136: performing time base alignment on the standardized environmental monitoring data set and the resource time change feature set according to the third association rule subset, performing sliding matching processing using an adaptive time window length, and generating a third association relationship subset.

[0061] The standardized environmental monitoring data set E_std and the resource time variation feature set T are time-aligned to ensure their temporal consistency. Then, a sliding matching method with adaptive time window length is used to match the two sets according to the third association rule subset R3. For each time window, the environmental monitoring indicators and resource time variation features within the window are checked to see if they meet the timing response constraints in the third association rule subset R3. If so, the association between the environmental monitoring indicators and resource time variation features within the time window is recorded to form the third association relationship subset A3.

[0062] Step S137: merging the first association relationship subset, the second association relationship subset, and the third association relationship subset into the resource dynamic association relationship set.

[0063] Merge the first association subset A1, the second association subset A2, and the third association subset A3 to form a resource dynamic association set A. This merging process can be performed using a concatenation method, combining corresponding elements from the three association subsets. Ultimately, the resource dynamic association set A = {A1, A2, A3} is obtained.

[0064] Step S140: generating a resource status evaluation strategy according to the resource dynamic association relationship set, and determining a resource management optimization direction based on the resource status evaluation strategy.

[0065] After obtaining the resource dynamic association relationship set A, it needs to be further analyzed and processed to generate a resource status assessment strategy and determine the optimization direction of resource management. The specific steps include the following:

[0066] Step S141: performing conflict detection processing on the resource dynamic association relationship set to identify a target association relationship subset having a conflicting relationship.

[0067] In order to ensure the accuracy and reliability of the resource dynamic association relationship set A, it is necessary to perform conflict detection. The specific steps are as follows:

[0068] Step S1411: obtaining a preset resource constraint condition set, wherein the resource constraint condition set includes ecological protection red line constraint conditions, resource carrying capacity constraint conditions, and land use planning constraint conditions.

[0069] The preset resource constraint set R_c contains multiple different types of constraints, which are based on relevant policies and regulations for natural resource management. Among them, ecological protection redline constraints are used to protect key ecological areas, resource carrying capacity constraints are used to ensure the rational use of resources, and land use planning constraints are used to regulate land development and utilization. These constraints are obtained from the preset resource constraint set.

[0070] Step S1412: performing matching verification processing on each association relationship subset in the resource dynamic association relationship set and the resource constraint condition set.

[0071] For each association relation subset in the resource dynamic association relation set A, it is matched and verified with each constraint condition in the resource constraint condition set R_c, and the association relation subset is checked to see whether it meets the requirements of each constraint condition.

[0072] Step S1413: If the association relationship subset conflicts with any constraint in the resource constraint condition set, the association relationship subset is marked as the target association relationship subset.

[0073] If a certain association relationship subset conflicts with any constraint in the resource constraint set R_c, such as violating the ecological protection red line constraint or exceeding the resource carrying capacity constraint, then the association relationship subset is marked as the target association relationship subset.

[0074] Step S1414: Count the conflict type distribution and conflict space distribution of the target association relationship subset and generate a conflict detection report.

[0075] Perform a statistical analysis of the relationships marked as the target relationship subset, calculating the conflict type distribution and conflict spatial distribution. The conflict type distribution reflects the frequency of occurrence of different types of conflicts within the target relationship subset, while the conflict spatial distribution shows the spatial location of conflicts within the target area. A conflict detection report is generated based on the statistical results, providing a basis for subsequent conflict resolution.

[0076] Step S142: calling a pre-trained decision tree model, performing priority sorting processing on the target association relationship subset, and generating a conflict resolution strategy set.

[0077] After identifying the target relationship subsets, in order to effectively resolve these conflicts, it is necessary to prioritize the target relationship subsets and determine a reasonable conflict resolution strategy. This step is achieved with the help of a pre-trained decision tree model. The specific process is as follows:

[0078] Step S1421: Acquire a historical conflict resolution case set, where the historical conflict resolution case set includes multiple historical conflict scenarios and corresponding solution priority tags.

[0079] Conflict resolution cases related to natural resource management in the current target area are collected from historical data to construct a historical conflict resolution case set. This historical conflict resolution case set consists of K cases, each consisting of a historical conflict scenario and a corresponding solution priority label. HCS_k denotes the kth historical conflict scenario, and SPL_k denotes the solution priority label corresponding to the kth case, where k ranges from 1 to K. Historical conflict scenarios can encompass different types of conflicts, such as those between ecological protection and resource development, and between land use planning and actual development. Solution priority labels rank the importance and urgency of different solutions based on historical experience.

[0080] Step S1422: performing feature extraction processing on the historical conflict resolution case set to generate a case feature set, wherein the case feature set includes conflict type features, spatial location features, cross-regional impact features, and resource impact degree features.

[0081] Feature extraction is performed for each historical conflict scenario (HCS_k) in the historical conflict resolution case collection. Conflict type features are categorized based on the nature and characteristics of the conflict, for example, into ecological conflicts, resource utilization conflicts, planning conflicts, and other types. CT_k represents the conflict type feature of the kth case.

[0082] The spatial location feature reflects the specific geographical location of the conflict, which is represented by geographic coordinates. Let the spatial location feature of the k-th case be SL_k.

[0083] The cross-regional impact feature measures whether a conflict will have an impact on surrounding areas. It is represented by a binary identifier, for example, 0 indicates no cross-regional impact and 1 indicates cross-regional impact. CRIF_k represents the cross-regional impact feature of the kth case.

[0084] The resource impact characteristic represents the degree of impact of the conflict on natural resources, which may involve the quantity, quality, sustainability, etc. RIF_k represents the resource impact characteristic of the kth case.

[0085] Combine these features to form the case feature set CF={CT_k, SL_k, CRIF_k, RIF_k|k=1, 2, ..., K}.

[0086] Step S1423: training the decision tree model based on the case feature set and the solution priority label, so that the decision tree model can predict the solution priority according to the input conflict feature.

[0087] Before training the decision tree model, the case feature set and solution priority labels need to be preprocessed. The specific steps are as follows:

[0088] For example, step S14231: performing one-hot encoding processing on the conflict type feature in the case feature set to generate a conflict type encoding vector, and performing plane coordinate conversion processing on the spatial position feature to generate a spatial position coordinate set in a unified coordinate system.

[0089] For the conflict type feature CT_k, use one-hot encoding to convert it into a vector. Assuming there are C conflict types, the conflict type feature CT_k for the kth case is converted into a vector of length C, where the position corresponding to the conflict type is 1 and the rest are 0. Let the converted conflict type encoding vector be CTE_k.

[0090] To ensure that all spatial location features of cases are processed in the same coordinate system, we transform them into a unified coordinate system using a plane coordinate transformation. Let the transformed spatial location coordinate set be SLC = {SLC_k | k = 1, 2, ..., K}.

[0091] Step S14232: performing binary identification conversion processing on the cross-regional impact feature to generate a cross-regional impact identification set.

[0092] Since the cross-regional impact feature CRIF_k itself is already in the form of a binary identifier, it is directly combined to form a cross-regional impact identifier set CRIF_set={CRIF_k|k=1, 2, ..., K}.

[0093] Step S14233: performing dimension normalization processing on the resource impact degree characteristics to generate a resource impact degree set of a unified interval.

[0094] The resource impact feature RIF_k may have different dimensions and value ranges. To eliminate the impact of dimensional differences on model training, it needs to be dimensionalized. Methods such as minimum-maximum normalization can be used to normalize the resource impact feature RIF_k to a uniform interval, such as [0, 1]. Let the normalized resource impact feature set be RIF_norm = {RIF_norm_k | k = 1, 2, ..., K}.

[0095] Step S14234: splicing the conflict type coding vector, spatial position coordinate set, cross-regional impact identification set and resource impact degree set according to the sample dimension to generate a dimensionally aligned normalized case feature set.

[0096] Concatenate the conflict type encoding vector CTE_k, the spatial location coordinate set SLC_k, the cross-regional impact identifier set CRIF_k, and the resource impact degree set RIF_norm_k according to the sample dimension. For the k-th case, concatenate its corresponding CTE_k, SLC_k, CRIF_k, and RIF_norm_k into a new vector. Combine the concatenated vectors of all cases to form a dimensionally aligned normalized case feature set NFCF = {NFCF_k | k = 1, 2, ..., K}.

[0097] Step S14235: Perform ordinal coding processing on the solution priority labels to generate an increasing priority label sequence.

[0098] Perform ordinal encoding on the solution priority labels SPL_k, sorting the different priority labels according to their importance and urgency, and encoding them with increasing integers. Let the encoded priority label sequence be SPL_encoded = {SPL_encoded_k | k = 1, 2, ..., K}.

[0099] Step S14236: performing recursive splitting calculation on the normalized case feature set and priority label sequence based on the splitting criterion to generate a decision tree node splitting rule set.

[0100] When training a decision tree model, the data is recursively split according to splitting criteria (such as information gain or the Gini index) based on the normalized case feature set NFCF and the priority label sequence SPL_encoded. For each node, a feature and a splitting threshold are selected to divide the data into two subsets, making the subsequent subsets more purified in terms of priority labels. This process is repeated until a stopping condition is met (such as when the number of samples in the node falls below a certain threshold or when the node purity reaches a certain threshold). During the splitting process, the splitting rule for each node is recorded to form the decision tree node splitting rule set NDRS.

[0101] Step S14237: Pruning the decision tree node splitting rule set according to preset pruning conditions, deleting redundant splitting nodes, and generating a simplified decision tree structure.

[0102] To prevent overfitting of the decision tree model, the generated decision tree node splitting rule set (NDRS) needs to be pruned. Preset pruning conditions can include maximum tree depth, minimum number of samples, and other factors. Based on these pruning conditions, redundant splitting nodes are removed, simplifying the decision tree structure. After pruning, a simplified decision tree structure (SDT) is generated.

[0103] Step S14238: Input the normalized case feature set into the simplified decision tree structure for training verification. If the classification accuracy does not meet the preset conditions, adjust the pruning conditions and retrain until the conditions are met to generate a trained decision tree model.

[0104] The normalized case feature set NFCF is input into the simplified decision tree structure SDT for training and validation. The classification accuracy is calculated based on the output prediction results and the actual priority label sequence SPL_encoded. If the classification accuracy does not meet the preset conditions (such as reaching a certain threshold), the pruning conditions are adjusted and the decision tree training and pruning process is repeated until the classification accuracy meets the preset conditions. Finally, a trained decision tree model DTM is generated.

[0105] Step S1424: inputting the conflict type distribution and conflict space distribution of the target association relationship subset into the trained decision tree model to obtain the solution priority of each target association relationship subset.

[0106] For each target relationship subset, extract the conflict type distribution and conflict spatial distribution features. These features are processed using the same preprocessing methods used when training the decision tree model, ensuring that they align with the feature dimensions and format of the normalized case feature set. The processed features are then fed into the trained decision tree model (DTM). The model outputs the solution priority for each target relationship subset. Assume there are L target relationship subsets. Let SPP_l represent the solution priority for the lth target relationship subset, where l ranges from 1 to L.

[0107] Step S1425: Sort the target association relationship subsets according to the solution priorities to generate a conflict resolution strategy set.

[0108] Sort the target association subsets according to the solution priority SPP_1 from high to low. Associate the sorted target association subsets with the corresponding solutions to form a conflict resolution strategy set CSS={CSS_1|l=1, 2, ..., L}, where CSS_1 represents the conflict resolution strategy corresponding to the lth target association subset.

[0109] Step S143: modifying the resource dynamic association relationship set based on the conflict resolution strategy set to generate an optimized resource dynamic association relationship set.

[0110] After obtaining the conflict resolution strategy set CSS, it is necessary to modify the resource dynamic association relationship set A to eliminate the conflicts. The specific steps are as follows:

[0111] Step S1431: Process the target association relationship subsets in descending order of priority of the solutions.

[0112] According to the priority of the solutions in the conflict resolution strategy set CSS, the target association subsets with the highest priority are processed in sequence. Let the target association subset currently being processed be the lth, and l starts at 1 and increases in sequence until all target association subsets are processed.

[0113] Step S1432: For each target association relationship subset, obtain its corresponding conflict type and conflict spatial position.

[0114] For the lth target association relationship subset, the corresponding conflict type CT_1 and conflict spatial location SL_1 are obtained from the conflict detection report. This information will be used to determine the specific correction rules.

[0115] Step S1433: calling a correction rule in a preset correction rule library according to the conflict type, wherein the correction rule library includes an ecological protection priority rule, a resource sustainable utilization rule, and a planning consistency rule.

[0116] The preset correction rule library contains various types of correction rules for handling different types of conflicts. Based on the conflict type CT_1 of the lth target association subset, the corresponding correction rule is invoked from the correction rule library. For example, if the conflict type is between ecological protection and resource development, the ecological protection priority rule is invoked; if the conflict is about resource sustainability, the resource sustainable utilization rule is invoked; and if the land use plan is inconsistent with the actual situation, the planning consistency rule is invoked.

[0117] Step S1434: performing logic adjustment or parameter adjustment on the target association relationship subset based on the correction rule to generate a corrected association relationship subset.

[0118] Based on the invoked correction rule, logic or parameter adjustments are made to the lth target association subset. Logical adjustments may involve changing the logical structure of the association, such as modifying certain conditional judgments; parameter adjustments may involve modifying parameters within the association, such as adjusting resource allocation ratios. After these adjustments, a corrected association subset A'_l is generated.

[0119] Step S1435: Replace the original association relationship subset with the revised association relationship subset, and update the resource dynamic association relationship set.

[0120] The revised association relationship subset A'_1 is used to replace the corresponding original association relationship subset in the resource dynamic association relationship set A. The updated resource dynamic association relationship set A is updated to a new set including the revised association relationship subset.

[0121] Step S1436: re-perform conflict detection on the updated resource dynamic association relationship set, and iteratively correct if new conflicts exist, until no conflicts are found, generating a final optimized resource dynamic association relationship set.

[0122] The updated set of dynamic resource associations is subjected to conflict detection again, following the procedure of step S141. If a new conflict is detected, steps S1431-S1435 are repeated for iterative correction until no more conflicts exist in the set of dynamic resource associations. This ultimately generates the optimized set of dynamic resource associations, A_optimized.

[0123] Step S144: performing segmented normalization processing on the optimized resource dynamic correlation relationship set according to index type differences to generate a standardized resource dynamic correlation relationship set eliminating dimensional influence, and generating a resource state evaluation index set according to the standardized resource dynamic correlation relationship set, the resource state evaluation index set including a resource utilization efficiency index, an ecological balance index, and a sustainability index, wherein, for the optimized resource dynamic correlation relationship set, the resource utilization efficiency index is normalized to the [0, 1] interval by using the minimum-maximum value normalization, the ecological balance index is standardized by using the Z-score standardization, and the sustainability index is normalized by using the quantile normalization.

[0124] In order to eliminate the dimensional differences between different indexes in the optimized resource dynamic correlation relationship set A_optimized and facilitate comprehensive evaluation, segmented normalization processing needs to be performed on it according to index type differences, as follows:

[0125] For the resource utilization efficiency index, the minimum-maximum value normalization method is used to normalize the value range of the index to the [0, 1] interval. Let the value of the resource utilization efficiency index in the optimized resource dynamic correlation relationship set A_optimized be RUE_i, where i represents different samples. RUE_i is converted into the normalized resource utilization efficiency index RUE_norm_i by the minimum-maximum value normalization formula.

[0126] For the ecological balance index, the distribution of the data needs to be checked before Z-score standardization is used. If the data conforms to the normal distribution, the Z-score standardization method can be directly used. Let the value of the ecological balance index in the optimized resource dynamic correlation relationship set A_optimized be EB_i, and the mean μ_EB and the standard deviation σ_EB are calculated. EB_i is converted into the standardized ecological balance index EB_std_i by the Z-score standardization formula. If the data distribution is skewed, the robust normalization method such as median-quartile scaling is used. The median M_EB and the interquartile range IQR_EB of the data are calculated, and each data point is subtracted from the median and divided by the interquartile range to obtain the robustly normalized ecological balance index.

[0127] For the sustainability index, when using the quantile normalization method, it is necessary to ensure that there is enough sample size. If the sample size is insufficient, the normalization result may not be accurate. In this case, the data quantity can be increased, or other normalization methods more suitable for small sample data can be used. The values of the sustainability index in the optimized resource dynamic correlation relationship set A_optimized are sorted, and they are converted into the normalized sustainability index SI_norm_i according to the definition of the quantile.

[0128] The normalized and standardized indicators are combined to form a standardized resource dynamic association set A_std that eliminates dimensionality effects. Based on A_std, the resource status assessment indicator set RSEI is generated, which includes the resource utilization efficiency indicator RUE_norm, the ecological balance indicator EB_std, and the sustainability indicator SI_norm.

[0129] Step S145: Determine the resource management optimization direction of the target area based on the resource status assessment indicator set, where the resource management optimization direction includes resource allocation adjustment direction, ecological protection priority direction, and monitoring network optimization direction.

[0130] Based on the indicators in the resource status evaluation index set RSEI, the resource management optimization direction of the target area is determined as follows:

[0131] The thresholds for resource efficiency, ecological balance, and sustainability indicators are not arbitrarily set; rather, they are dynamically adjusted based on historical data or expert knowledge. By analyzing historical data, statistically analyzing the ranges and distributions of values ​​for resource efficiency, ecological balance, and sustainability indicators over different time periods, and incorporating expert experience and advice, appropriate threshold ranges are determined. Furthermore, to ensure the rationality and stability of the thresholds, sensitivity analysis is also required. This involves varying the threshold values, observing changes in the direction of resource management optimization, and evaluating the extent to which the thresholds affect the results.

[0132] Step S1451: If the resource utilization efficiency index after segment normalization is lower than a preset first threshold, a resource allocation adjustment direction is generated, which includes optimizing resource mining layout, adjusting resource allocation ratio and introducing efficient utilization technology.

[0133] The segmented, normalized resource utilization efficiency indicator, RUE_norm, is compared with a first threshold dynamically adjusted based on historical data and expert knowledge. If RUE_norm is lower than the first threshold, resource utilization efficiency in the target area is low, and resource allocation adjustments are necessary. These adjustments include optimizing resource extraction layout, such as replanning extraction areas to avoid excessive concentration and waste; adjusting resource allocation ratios to rationally allocate resources based on the needs of different regions and industries; and introducing high-efficiency utilization technologies to improve resource utilization efficiency.

[0134] Step S1452: If the segmented normalized ecological balance index is lower than the preset second threshold, an ecological protection priority direction is generated, which includes demarcating ecological restoration areas, strengthening ecological monitoring frequency, and limiting the intensity of development activities.

[0135] The normalized ecological balance indicator (EB_std) is compared with a second threshold dynamically adjusted based on historical data and expert knowledge. If EB_std falls below the second threshold, it indicates that the ecological balance in the target area has been damaged to a certain extent, and ecological protection priorities need to be determined. These priorities include designating ecological restoration areas to repair damaged ecological areas; increasing the frequency of ecological monitoring to keep abreast of changes in the ecological environment; and limiting the intensity of development activities to reduce further damage to the ecological environment.

[0136] Step S1453: If the sustainability index after segment normalization is lower than the preset third threshold, a monitoring network optimization direction is generated, which includes increasing the density of monitoring points, optimizing the combination of monitoring parameters, and upgrading the accuracy of monitoring equipment.

[0137] The segmented, normalized sustainability indicator, SI_norm, is compared with a third threshold dynamically adjusted based on historical data and expert knowledge. If SI_norm falls below the third threshold, it indicates resource sustainability issues in the target area, necessitating optimization of the monitoring network. This optimization approach includes increasing the density of monitoring points to improve resource and environmental monitoring coverage; optimizing monitoring parameter combinations to ensure that monitoring data accurately reflects changes in resources and the environment; and upgrading monitoring equipment to enhance the accuracy and reliability of monitoring data.

[0138] Step S150: Feedback the resource management optimization direction to the resource management system to trigger resource scheduling operations.

[0139] After determining the resource management optimization direction, these directions need to be converted into specific instructions and fed back to the resource management system to trigger resource scheduling operations. The specific steps are as follows:

[0140] Step S151: converting the resource allocation adjustment direction into a resource scheduling instruction set, wherein the resource scheduling instruction set includes mining equipment scheduling instructions, transportation route adjustment instructions, and storage facility optimization instructions.

[0141] For resource allocation adjustments, they are converted into a specific set of resource scheduling instructions. For optimizing resource extraction layout, mining equipment scheduling instructions are generated, such as relocating mining equipment to a more suitable mining area. For adjusting resource allocation ratios, transportation route adjustment instructions are generated to ensure that resources can be transported according to the new allocation ratio. For introducing efficient utilization technologies, storage facility optimization instructions are generated, such as upgrading and renovating storage facilities to meet the requirements of efficient utilization technologies. Let the resource scheduling instruction set be RSI = {RSI_1, RSI_2, RSI_3}, where RSI_1 represents mining equipment scheduling instructions, RSI_2 represents transportation route adjustment instructions, and RSI_3 represents storage facility optimization instructions.

[0142] Step S152: converting the ecological protection priority direction into an ecological control instruction set, wherein the ecological control instruction set includes an area entry ban instruction, an activity approval instruction, and a restoration project start instruction.

[0143] For ecological protection priorities, these are converted into a set of ecological control instructions. For demarcating ecological restoration areas, regional entry prohibitions are generated, prohibiting unauthorized personnel and activities from entering the restoration area. For strengthening ecological monitoring frequency, activity approval instructions are generated, strictly approving activities that may impact the ecological environment. For limiting the intensity of development activities, restoration project initiation instructions are generated to initiate restoration projects in damaged ecological areas. Let the set of ecological control instructions be ECI = {ECI_1, ECI_2, ECI_3}, where ECI_1 represents the regional entry prohibition instruction, ECI_2 represents the activity approval instruction, and ECI_3 represents the restoration project initiation instruction.

[0144] Step S153: converting the monitoring network optimization direction into a monitoring equipment configuration instruction set, wherein the monitoring equipment configuration instruction set includes an equipment purchase instruction, an equipment installation location instruction, and a data collection frequency adjustment instruction.

[0145] For monitoring network optimization, this is converted into a set of monitoring equipment configuration instructions. For increasing monitoring point density, equipment procurement instructions are generated to procure the appropriate monitoring equipment based on the number and type of monitoring points required. Assuming the number of equipment to be procured is M, with n_1, n_2, ..., n_M of each type, we analyze the monitoring needs of the target area and combine them with the existing distribution of monitoring points to determine the specific equipment quantity and type, generating equipment procurement instructions.

[0146] To optimize the combination of monitoring parameters, device installation location instructions are generated. Different monitoring parameters may require collection at different locations to obtain more accurate data. For example, air quality monitoring may require monitoring points at different altitudes and in different areas. By analyzing factors such as the target area's geographical environment and meteorological conditions, the optimal installation location for each monitoring device is determined, and device installation location instructions are generated.

[0147] To upgrade the accuracy of monitoring equipment, generate data collection frequency adjustment instructions. The upgraded monitoring equipment may have higher accuracy and faster data collection capabilities, requiring corresponding adjustments to the data collection frequency. Based on the performance of the monitoring equipment and changes in monitoring indicators, determine the appropriate data collection frequency and generate data collection frequency adjustment instructions. Let the set of monitoring equipment configuration instructions be MDI = {MDI_1, MDI_2, MDI_3}, where MDI_1 represents the equipment purchase instruction, MDI_2 represents the equipment installation location instruction, and MDI_3 represents the data collection frequency adjustment instruction.

[0148] Step S154: Based on the spatial topology relationship verification rules and instruction priority rules, perform spatial conflict detection and logical coordination processing on the resource scheduling instruction set, ecological management and control instruction set, and monitoring equipment configuration instruction set, and send the coordinated instruction set to the resource management system to trigger a conflict-free resource scheduling operation.

[0149] Before sending the resource scheduling instruction set RSI, the ecological control instruction set ECI, and the monitoring device configuration instruction set MDI to the resource management system, spatial conflict detection and logical coordination processing are required to ensure that there are no conflicts between these instructions. The details are as follows:

[0150] First, the spatial topology validation rules were expanded to cover interactions across all instruction types (resource scheduling, ecological control, and monitoring equipment). In addition to considering overlaps between mining and restricted areas, conflicts between the installation location of monitoring equipment and transportation routes also needed to be considered. For example, would the equipment obstruct transportation? Would the location of storage facilities in resource scheduling instructions affect the normal operation of monitoring equipment? Would activity approvals in ecological control instructions conflict with mining activities in resource scheduling instructions?

[0151] For spatial topological relationship verification rules, consider the spatial relationships between various instructions. For example, mining equipment scheduling instructions within resource scheduling instructions may involve specific mining areas, while regional prohibition instructions within ecological management instructions may restrict certain areas. It is necessary to check whether the areas covered by mining equipment scheduling instructions overlap with those restricted by regional prohibition instructions. Let A_RSI_1 be the set of areas covered by mining equipment scheduling instructions, and A_ECI_1 be the set of areas restricted by regional prohibition instructions. By performing spatial analysis on these two area sets, determine whether they intersect. If A_RSI_1 ∩ A_ECI_1 ≠ ∅, then a spatial conflict exists.

[0152] At the same time, check whether the installation location determined by the monitoring device installation location instruction conflicts with the transportation route planned by the transportation route adjustment instruction. Let A_MDI_2 be the set of monitoring device installation locations, and A_RSI_2 be the set of areas covered by the transportation route. If A_MDI_2 ∩ A_RSI_2 ≠ ∅, then the monitoring device may obstruct transportation, indicating a spatial conflict.

[0153] Logical coordination requires compliance with directive priority rules. These rules are based on the importance and urgency of resource management. For example, directives related to ecological protection typically have a higher priority because ecological protection is crucial to sustainable development.

[0154] Assuming that the Ecological Control Instruction Set (ECI) has a higher priority than the Resource Scheduling Instruction Set (RSI), when spatial conflicts arise, the Resource Scheduling Instructions need to be adjusted. If the areas covered by the mining equipment scheduling instructions overlap with the areas restricted by the regional prohibition instructions, the mining equipment scheduling plan needs to be replanned to avoid entering the restricted area. If the installation location of the monitoring equipment conflicts with the transportation route, and the Ecological Control Instruction Set (ECI) has a higher priority, the installation location of the monitoring equipment needs to be readjusted to ensure that transportation is not affected.

[0155] The transport route adjustment instructions and storage facility optimization instructions in the resource scheduling instruction set (RSI) also need to be logically coordinated with the ecological control instruction set (ECI) and the monitoring equipment configuration instruction set (MDI). For example, the transport route planned by the transport route adjustment instruction must not damage the ecological restoration area, and the storage facility construction location involved in the storage facility optimization instruction must not affect the normal operation of the monitoring equipment.

[0156] Similarly, for the installation location instructions of the monitoring device in the monitoring device configuration instruction set MDI, it is necessary to ensure that the installation location does not conflict with the resource scheduling instructions and the ecological management instructions. For example, the installation location of the monitoring device cannot affect the normal operation of the mining device, nor can it violate the regional access restriction instructions. When checking the relationship between the storage facility construction location and the monitoring device, set the storage facility construction location set as A_RSI_3, if A_RSI_3 has an intersection with the monitoring device installation location set A_MDI_2, that is, A_RSI_3∩A_MDI_2≠∅, it may affect the normal operation of the monitoring device, at which time the construction location of the storage facility needs to be re-planned.

[0157] For the activity approval instructions in the ecological management instruction set ECI, logical coordination is required with the mining activities, transportation activities, etc. in the resource scheduling instruction set RSI. If the mining activities or transportation activities do not meet the requirements of the activity approval instructions, for example, they may cause great damage to the ecological environment, then these activities need to be suspended or adjusted until the approval conditions are met.

[0158] After the spatial conflict detection and logical coordination processing, the coordinated instruction set is sent to the resource management system. After receiving the coordinated instruction set, the resource management system will trigger corresponding resource scheduling operations according to these instructions. For example, according to the mining device scheduling instructions in the resource scheduling instruction set RSI, the mining devices are allocated, according to the transportation path adjustment instructions, the transportation routes are adjusted, and according to the storage facility optimization instructions, the storage facilities are optimized; according to the regional access restriction instructions in the ecological management instruction set ECI, the regional restrictions are set, according to the activity approval instructions, the related activities are approved, and according to the repair engineering start instructions, the ecological repair engineering is started; according to the device procurement instructions in the monitoring device configuration instruction set MDI, the monitoring devices are procured, according to the device installation location instructions, the devices are installed, and according to the data collection frequency adjustment instructions, the data collection frequency is adjusted.

[0159] During the entire process, it is also necessary to continuously monitor and feedback the execution of the resource scheduling operations. By collecting the data of various operations in the resource management system in real time, a comparative analysis is made with the expected resource management optimization target. If it is found that the actual execution deviates from the expected target, for example, the resource utilization efficiency has not been effectively improved, the ecological balance index is still below the threshold, or the sustainability index has not been significantly improved, it is necessary to re-examine the previous analysis process and instruction generation process. It may be necessary to check the accuracy and completeness of the multi-source data again, adjust the dynamic association rule set, optimize the resource state evaluation strategy, or re-determine the resource management optimization direction and the corresponding instruction set, so as to ensure that the natural resources in the target area can be managed and utilized scientifically, reasonably and effectively, and gradually achieve the sustainable development of resources and the balanced protection of the ecological environment.

[0160] Meanwhile, as time goes by and new data is accumulated, the natural resource analysis method combining multi-source data as a whole needs to be continuously optimized. For example, periodically update the pre-trained semantic segmentation model, time series analysis model, and decision tree model, etc. to improve the accuracy and adaptability of the model. Dynamically adjust the dynamic association rule set so that it can better reflect the internal relationship and change rule between natural resources. Continuously improve the spatial topological relationship verification rules and instruction priority rules to cope with more complex resource scheduling and management scenarios.

[0161] Figure 2 A schematic diagram showing exemplary hardware and software components of the natural resource analysis system 100 combining multi-source data according to some embodiments of the present application is shown. For example, the processor 120 can be used in the natural resource analysis system 100 combining multi-source data and used to perform the functions in the present application.

[0162] The natural resource analysis system 100 combining multi-source data can be a general-purpose server or a special-purpose server, both of which can be used to implement the natural resource analysis method combining multi-source data of the present application. Although only one server is shown in the present application, for the sake of convenience, the functions described in the present application can be implemented in a distributed manner on multiple similar platforms to balance the processing load.

[0163] For example, the natural resource analysis system 100 combining multi-source data can include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and different forms of storage media 140, such as a disk, a ROM, or a RAM, or any combination thereof. Exemplarily, the natural resource analysis system 100 combining multi-source data can also include program instructions stored in a ROM, a RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present application can be implemented according to these program instructions. The natural resource analysis system 100 combining multi-source data also includes an I / O interface 150 between the computer and other input / output devices.

[0164] For ease of explanation, only one processor is described in the natural resource analysis system 100 that combines multi-source data. However, it should be noted that the natural resource analysis system 100 that combines multi-source data in the present invention may also include multiple processors, so the steps performed by one processor described in the present invention may also be performed jointly or individually by multiple processors. For example, if the processor of the natural resource analysis system 100 that combines multi-source data executes step A and step B, it should be understood that step A and step B may also be performed jointly by two different processors or individually in one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor execute steps A and B together.

[0165] In addition, an embodiment of the present invention further provides a readable storage medium, in which computer-executable instructions are preset. When a processor executes the computer-executable instructions, the above-mentioned natural resource analysis method combining multi-source data is implemented.

[0166] It should be noted that in order to simplify the description of the present invention and thus help understand one or more embodiments of the invention, in the foregoing description of the embodiments of the present invention, multiple features are sometimes combined into one embodiment, figure or description thereof.

Claims

1. A natural resource analysis method combining multi-source data, characterized in that: The method comprises: Acquire a multi-source data set formed by multiple data sources of a target area, wherein the multi-source data set includes a remote sensing image data set, a geographic monitoring data set, and an environmental monitoring data set; Performing spatiotemporal feature extraction processing on the multi-source data set to obtain a resource spatial distribution feature set and a resource temporal change feature set of the target area; Based on a preset dynamic association rule set, dynamically associate the resource spatial distribution feature set and the resource time change feature set to generate a resource dynamic association relationship set; Generating a resource status evaluation strategy according to the resource dynamic association relationship set, and determining a resource management optimization direction based on the resource status evaluation strategy; Feedback of the resource management optimization direction to the resource management system to trigger resource scheduling operations; The performing spatiotemporal feature extraction processing on the multi-source data set to obtain a resource spatial distribution feature set and a resource temporal change feature set of the target area includes: Performing spatial resolution alignment processing and unit uniform conversion processing on the remote sensing image data set to generate a standardized remote sensing image data set; Calling a pre-trained semantic segmentation model to perform surface cover type recognition processing on the standardized remote sensing image data set to obtain a surface cover distribution feature set of the target area; Performing spatial reference coordinate system conversion and unit unification processing on the geographic monitoring data set to generate a terrain elevation distribution feature set and a soil type distribution feature set consistent with the resolution of the standardized remote sensing image; Performing time series alignment and granularity unification processing on the environmental monitoring data set to generate a standardized environmental monitoring data set; Calling a pre-trained time series analysis model to perform periodic change pattern extraction processing on the standardized environmental monitoring data set to obtain a resource time change feature set in the target area; Merging the land cover distribution feature set, the terrain elevation distribution feature set, and the soil type distribution feature set into the resource space distribution feature set; The dynamically associating the resource spatial distribution feature set and the resource temporal change feature set based on a preset dynamic association rule set to generate a resource dynamic association relationship set includes: Acquire a first association rule subset in the dynamic association rule set, where the first association rule subset is used to characterize a spatial association constraint condition between a land cover type and a terrain elevation; Performing resolution consistency check on the land cover distribution feature set and the terrain elevation distribution feature set according to the first association rule subset, performing spatial matching processing in a unified grid coordinate system, and generating a first association relationship subset; Acquire a second association rule subset in the dynamic association rule set, where the second association rule subset is used to characterize a spatial compatibility constraint condition between soil types and land cover types; Performing spatial overlay analysis on the soil type distribution feature set and the land cover distribution feature set after unit unification according to the second association rule subset to generate a second association relationship subset; Acquire a third association rule subset in the dynamic association rule set, where the third association rule subset is used to characterize a temporal response constraint condition between an environmental monitoring indicator and a resource time variation characteristic; Performing time base alignment on the standardized environmental monitoring data set and the resource time change feature set according to the third association rule subset, performing sliding matching processing using an adaptive time window length, and generating a third association relationship subset; The first association relationship subset, the second association relationship subset, and the third association relationship subset are merged into the resource dynamic association relationship set.

2. The natural resource analysis method combining multi-source data according to claim 1, characterized in that: Generating a resource status evaluation strategy according to the resource dynamic association relationship set, and determining a resource management optimization direction based on the resource status evaluation strategy, includes: Performing conflict detection on the resource dynamic association relationship set to identify a target association relationship subset with conflicting relationships; Calling a pre-trained decision tree model to prioritize the target association relationship subset and generate a conflict resolution strategy set; Modifying the resource dynamic association relationship set based on the conflict resolution strategy set to generate an optimized resource dynamic association relationship set; Performing piecewise normalization processing on the optimized resource dynamic association relationship set according to indicator type differences to generate a standardized resource dynamic association relationship set that eliminates dimensionality effects, and generating a resource status assessment indicator set based on the standardized resource dynamic association relationship set, wherein the resource status assessment indicator set includes a resource utilization efficiency indicator, an ecological balance indicator, and a sustainability indicator, wherein the optimized resource dynamic association relationship set is normalized to the interval [0, 1] using the minimum-maximum value according to the resource utilization efficiency indicator, the ecological balance indicator is normalized using the Z-score, and the sustainability indicator is normalized using the quantile; The resource management optimization direction of the target area is determined based on the resource status evaluation indicator set, and the resource management optimization direction includes resource allocation adjustment direction, ecological protection priority direction and monitoring network optimization direction.

3. The natural resource analysis method combining multi-source data according to claim 2, characterized in that: The performing conflict detection on the resource dynamic association relationship set to identify a target association relationship subset having a conflicting relationship includes: Obtaining a preset resource constraint condition set, wherein the resource constraint condition set includes ecological protection red line constraint conditions, resource carrying capacity constraint conditions, and land use planning constraint conditions; Performing matching verification processing on each association relationship subset in the resource dynamic association relationship set and the resource constraint condition set; If the association relationship subset conflicts with any constraint in the resource constraint condition set, marking the association relationship subset as the target association relationship subset; The conflict type distribution and conflict space distribution of the target association relationship subset are counted to generate a conflict detection report.

4. The natural resource analysis method combining multi-source data according to claim 2, characterized in that: The calling of the pre-trained decision tree model, performing priority sorting on the target association relationship subset, and generating a conflict resolution strategy set includes: Acquire a historical conflict resolution case set, where the historical conflict resolution case set includes multiple historical conflict scenarios and corresponding resolution priority labels; Performing feature extraction processing on the historical conflict resolution case set to generate a case feature set, wherein the case feature set includes conflict type features, spatial location features, cross-regional impact features, and resource impact degree features; Training the decision tree model based on the case feature set and the solution priority label, so that the decision tree model can predict the solution priority according to the input conflict feature; Inputting the conflict type distribution and conflict spatial distribution of the target association relationship subset into the trained decision tree model to obtain a solution priority for each target association relationship subset; The target association relationship subsets are sorted according to the solution priorities to generate a conflict resolution strategy set.

5. The natural resource analysis method combining multi-source data according to claim 4 is characterized in that: The modifying the resource dynamic association relationship set based on the conflict resolution strategy set to generate an optimized resource dynamic association relationship set includes: Processing the target association relationship subsets in descending order of priority of the solutions; For each target association relationship subset, obtain its corresponding conflict type and conflict spatial location; Invoking a correction rule in a preset correction rule library according to the conflict type, wherein the correction rule library includes an ecological protection priority rule, a resource sustainable utilization rule, and a planning consistency rule; Performing logic adjustment or parameter adjustment on the target association relationship subset based on the correction rule to generate a corrected association relationship subset; Replacing the original association relationship subset with the revised association relationship subset, and updating the resource dynamic association relationship set; The updated resource dynamic association relationship set is re-checked for conflicts. If new conflicts exist, they are iteratively corrected until no conflicts are found, generating the final optimized resource dynamic association relationship set.

6. The natural resource analysis method combining multi-source data according to claim 2, characterized in that: The determining of the resource management optimization direction of the target area based on the resource status evaluation indicator set includes: If the resource utilization efficiency index after segment normalization is lower than a preset first threshold, a resource allocation adjustment direction is generated, wherein the resource allocation adjustment direction includes optimizing resource mining layout, adjusting resource allocation ratio, and introducing efficient utilization technology; If the segmented normalized ecological balance index is lower than a preset second threshold, an ecological protection priority direction is generated, which includes demarcating ecological restoration areas, strengthening ecological monitoring frequency, and limiting the intensity of development activities; If the sustainability index after segmented normalization is lower than a preset third threshold, a monitoring network optimization direction is generated, which includes increasing the density of monitoring points, optimizing the combination of monitoring parameters, and upgrading the accuracy of monitoring equipment.

7. The natural resource analysis method combining multi-source data according to claim 6, characterized in that: Feedback of the resource management optimization direction to the resource management system to trigger resource scheduling operations includes: Converting the resource allocation adjustment direction into a resource scheduling instruction set, wherein the resource scheduling instruction set includes mining equipment scheduling instructions, transportation route adjustment instructions, and storage facility optimization instructions; Converting the ecological protection priority direction into an ecological control instruction set, wherein the ecological control instruction set includes an area entry ban instruction, an activity approval instruction, and a restoration project start instruction; Converting the monitoring network optimization direction into a monitoring device configuration instruction set, wherein the monitoring device configuration instruction set includes an equipment purchase instruction, an equipment installation location instruction, and a data acquisition frequency adjustment instruction; Based on the spatial topology relationship verification rules and instruction priority rules, spatial conflict detection and logical coordination processing are performed on the resource scheduling instruction set, ecological management and control instruction set and monitoring equipment configuration instruction set, and the coordinated instruction set is sent to the resource management system to trigger conflict-free resource scheduling operations.

8. A natural resource analysis system combining multi-source data, characterized in that: It includes a processor and a memory, the memory is connected to the processor, the memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to implement the natural resource analysis method combining multi-source data as described in any one of claims 1 to 7 above.

Images