Risk diagnosis method and device for activity layer thickness spatial mapping, storage medium and electronic equipment

Abstract

The application provides a risk diagnosis method and device for activity layer thickness spatial mapping, a storage medium and an electronic device, and relates to geographic spatial information. The electronic device obtains environmental information of a plurality of first sub-regions in a region to be analyzed, and then processes the environmental information of each sub-region by using a pre-trained thickness prediction model to obtain corresponding predicted activity layer thickness. Then, the measured activity layer thickness of a plurality of reference sub-regions having measured data is compared with the model prediction value one by one, and the thickness reliability of the plurality of reference sub-regions is calculated. Then, the thickness reliability of the remaining first sub-regions without measured data is calculated according to the thickness reliability. In this way, the user can obtain the activity layer thickness distribution map and intuitively identify which region's prediction result needs to be used with caution based on the thickness reliability.

Description

Technical Field

[0001] This application relates to the field of geospatial information, and more specifically, to a risk diagnosis method, apparatus, storage medium, and electronic device for spatial mapping of active layer thickness. Background Technology

[0002] Active layer thickness is a crucial parameter characterizing the surface thermal state, hydrothermal migration processes, land-atmosphere energy exchange, and permafrost stability in permafrost regions. It is also a key fundamental variable in permafrost degradation assessment, ecological environment response analysis, and regional environmental simulation. Because active layer thickness is influenced by multiple factors such as topography, vegetation cover, soil properties, snow cover, and local climate, its spatial distribution typically exhibits significant local variability and strong environmental heterogeneity. Therefore, obtaining continuous distribution results of active layer thickness at the regional scale has always been an important technical requirement in permafrost research.

[0003] However, existing active layer thickness mapping often focuses on the model fitting accuracy or the mapping results themselves, lacking effective guidance on the reliability of regional extrapolation, and is particularly difficult to identify high-risk prediction areas in vegetation boundaries, valley transition zones and sparse sample areas. Summary of the Invention

[0004] In order to overcome at least one of the shortcomings of the prior art, one of the objectives of this application is to provide a risk diagnosis method, apparatus, storage medium and electronic device for spatial mapping of active layer thickness, which enables users to obtain an active layer thickness distribution map and intuitively identify which areas' prediction results should be used with caution based on thickness confidence or thickness confidence.

[0005] In a first aspect, this application provides a risk diagnosis method for spatial mapping of active layer thickness, the method comprising: Obtain the environmental information of each of the multiple first sub-regions in the region to be analyzed; The environmental information of each first sub-region is processed using a pre-trained thickness prediction model to obtain the predicted active layer thickness of each first sub-region. The measured active layer thickness of each of the multiple reference sub-regions is compared with the corresponding predicted active layer thickness to obtain the thickness confidence of the multiple reference sub-regions, wherein each of the reference sub-regions represents a sub-region with a measured active layer thickness; Based on the thickness confidence of the plurality of reference sub-regions, the thickness confidence of the remaining sub-regions in the plurality of first sub-regions is obtained.

[0006] Secondly, this application provides a risk diagnosis device for spatial mapping of active layer thickness, the device comprising: The environmental information module is used to obtain the environmental information of each of the multiple first sub-regions in the area to be analyzed; The thickness prediction module is used to process the environmental information of each of the first sub-regions using a pre-trained thickness prediction model to obtain the predicted active layer thickness of each of the first sub-regions. The risk prediction module is used to compare the measured thickness of the active layer of each of the multiple reference sub-regions with the corresponding predicted thickness of the active layer to obtain the thickness confidence of the multiple reference sub-regions, wherein each of the reference sub-regions represents a sub-region with a measured thickness of the active layer; The risk prediction module is further configured to obtain the thickness confidence of the remaining sub-regions in the plurality of first sub-regions based on the thickness confidence of the plurality of reference sub-regions.

[0007] Thirdly, this application provides a storage medium storing a computer program that, when executed by a processor, implements the risk diagnosis method for spatial mapping of active layer thickness.

[0008] Fourthly, this application provides an electronic device, which includes a processor and a memory. The memory stores a computer program, which, when executed by the processor, implements the risk diagnosis method for spatial mapping of active layer thickness.

[0009] Compared with the prior art, this application has the following beneficial effects: The risk diagnosis method, apparatus, storage medium, and electronic device for spatial mapping of active layer thickness provided in this application involve the electronic device acquiring environmental information of multiple first sub-regions within the area to be analyzed. A pre-trained thickness prediction model is then used to process the environmental information of each sub-region to obtain the corresponding predicted active layer thickness. Subsequently, the measured active layer thickness of multiple reference sub-regions with actual measurement data is compared one by one with the model prediction values ​​to calculate the thickness confidence level of each reference sub-region. Based on these thickness confidence levels, the thickness confidence level of the remaining first sub-regions without actual measurement data is then deduced. In this way, while obtaining the active layer thickness distribution map, users can intuitively identify which regions' prediction results require caution based on the thickness confidence level. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1One of the flowcharts for a risk diagnosis method for spatial mapping of active layer thickness provided in this application embodiment; Figure 2 A second schematic flowchart of the risk diagnosis method for spatial mapping of active layer thickness provided in this application embodiment; Figure 3 A schematic diagram of the structure of the risk diagnosis device for spatial mapping of active layer thickness provided in this application embodiment; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of this application (hereinafter referred to as "the embodiments") clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0013] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0014] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0015] In the description of this application, it should be noted that the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0016] Based on the above statement, as introduced in the background section, existing active layer thickness mapping often focuses on the model fitting accuracy or the mapping result itself, and lacks an effective expression of the reliability in regional extrapolation.

[0017] Specifically, relevant technologies for obtaining active layer thickness mainly include methods such as pit observation, steel rod detection, shallow drilling, and temperature probe observation. While these methods can obtain relatively reliable observation results at the point scale, they typically suffer from high field costs, limited spatial coverage, insufficient sample numbers, and weak ability to represent regional continuity, especially in complex terrain conditions such as high-altitude cold mountainous areas, slope-valley transition zones, and permafrost watersheds. Relying solely on sparse point samples is usually insufficient to effectively support regional-scale spatial mapping of active layer thickness and identification of high-risk areas.

[0018] In light of this, a further proposed technique utilizes Ground Penetrating Radar (GPR) to rapidly identify subsurface reflective interfaces along continuous profiles, providing high-density, continuous observational information for active layer thickness surveys. Simultaneously, digital elevation models and their derived topographic factors, climate factors, snow cover factors, vegetation factors, and soil property factors—multi-source raster data—can reflect the environmental control context of active layer thickness from different perspectives. Unifying the organization of GPR-derived samples, conventional point samples, and multi-source environmental factors, and employing machine learning methods for regional estimation, is a crucial technical approach to improving the accuracy and efficiency of active layer thickness mapping in permafrost regions.

[0019] However, when mapping the active layer thickness, related technologies often focus on the model fitting accuracy or the mapping result itself, and lack an effective expression of the reliability in regional extrapolation. This makes it difficult for users to perceive the reliability of the active layer thickness in the mapping.

[0020] It should be noted that the defects in the solutions in the prior art are the result of practice and careful research. Therefore, the discovery process of the above problems and the solutions proposed by the embodiments of this application in the following text should be regarded as contributions to this application in the process of invention and creation, and should not be understood as technical content known to those skilled in the art.

[0021] Based on the discovery of the above-mentioned technical problems, this embodiment provides a risk diagnosis method for spatial mapping of active layer thickness. For example... Figure 1 As shown, the method includes: S1, obtain the environmental information of each of the multiple first sub-regions in the area to be analyzed; S2, using a pre-trained thickness prediction model to process the environmental information of each first sub-region, to obtain the predicted thickness of the active layer of each first sub-region; S3 compares the measured thickness of the active layer of each of the multiple reference sub-regions with the predicted thickness of the corresponding active layer to obtain the thickness confidence of the multiple reference sub-regions.

[0022] Each reference sub-region represents a sub-region with the measured thickness of the active layer; S4. Based on the thickness confidence of multiple reference sub-regions, obtain the thickness confidence of the remaining sub-regions in the multiple first sub-regions.

[0023] This embodiment can be understood as follows: It acquires environmental information for each of the multiple first sub-regions within the area to be analyzed, then processes this information using a pre-trained thickness prediction model to obtain the corresponding predicted active layer thickness. Subsequently, it compares the measured active layer thickness of multiple reference sub-regions with actual data with the model's predicted values ​​one by one, calculating the thickness confidence level of each reference sub-region. Based on these thickness confidence levels, the thickness confidence level of the remaining first sub-regions without actual data is then deduced. In this way, while obtaining the active layer thickness distribution map, users can intuitively identify which regions' prediction results require caution based on the thickness confidence level.

[0024] It should be noted that the risk diagnosis method for activity layer thickness spatial mapping provided in this embodiment can be implemented using electronic devices such as mobile terminals, tablet computers, laptop computers, desktop computers, and servers. The server can be a single server or a group of servers. The server group can be centralized or distributed (e.g., the server can be a distributed system). In some embodiments, the server can be local or remote relative to the user terminal. In some embodiments, the server can be implemented on a cloud platform; as an example only, the cloud platform can include private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, inter-cloud, multi-cloud, etc., or any combination thereof. In some embodiments, the server can be implemented on an electronic device having one or more components.

[0025] To make the solution provided in this embodiment clearer, a server is used as the electronic device for implementing the method below, and in conjunction with... Figure 1 Each step of the method is described in detail. However, it should be understood that the operations in the flowchart may not be implemented in sequence, and steps without logical contextual relationships may be reversed in order or implemented simultaneously. Furthermore, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowchart, or remove one or more operations from the flowchart. Figure 1 As shown, the method includes: S1, obtain the environmental information of each of the multiple first sub-regions in the region to be analyzed.

[0026] In this embodiment, the server can perform multi-source heterogeneous organization and standardization processing on the geospatial data of the region to be analyzed to obtain environmental information corresponding to each first sub-region.

[0027] In practical applications, the Wenquan area in the northeastern part of the Qinghai-Tibet Plateau was selected as the region to be analyzed. This region belongs to the transition zone between permafrost and seasonal permafrost, featuring a landform pattern of alternating high-altitude mountains, intermontane basins, and valleys. The surface environment exhibits strong heterogeneity, with an altitude spanning 2000–5000 m, and significant vegetation zonation and complex topographic undulations. Therefore, it is suitable for verifying the risk diagnosis method for spatial mapping of active layer thickness provided in this embodiment. Based on this, the server can use 30 m as the preset analysis grid resolution to divide the entire region to be analyzed into several first sub-regions, each corresponding to a 30 m × 30 m raster pixel.

[0028] It should also be understood that this environmental information is a set of multi-source environmental factors constructed to characterize the spatial differences in the thickness of the active layer, including different dimensions such as climate, topography, surface and soil, and snow cover. Therefore, the server can acquire climate factors including annual temperature (AT), annual precipitation (Pre), land surface temperature (LST), thawing degree days (TDD), and freezing degree days (FDD); topographic factors including longitude (Lon), latitude (Lat), digital elevation model (DEM), slope, aspect, topographic wetness index (TWI), and potential solar radiation; normalized difference vegetation index (NDVI), normalized difference water index (NDWI), and normalized difference moisture index (NDMI); and bulk density (BD) and sand content. Soil factors including soil content, clay content, and cobble fraction (Cf); and snow cover factors including snow cover duration (SCD).

[0029] In this process, the server can uniformly reproject all environmental factors to the Albers iso-area coordinate system and unify the 1 km resolution climate product and the 500 m resolution snow cover product into a 30 m analysis grid through resampling to balance regional background constraints and local detail representation. Simultaneously, the server can use Spearman correlation analysis to screen candidate factors for collinearity. When the absolute value of the correlation coefficient between two factors exceeds a preset threshold, factors with clearer physical meaning or higher measured reliability are prioritized for retention. For example, in the case of a high negative correlation between soil organic carbon density (Soc) and bulk density (BD), Soc is removed while BD is retained. Finally, the server can accurately extract the corresponding pixel values ​​from the uniformly processed environmental factor raster based on the center coordinates of each first sub-region, and stitch all the extracted values ​​together in a fixed order to form environmental information specific to each first sub-region, with complete dimensions and clear semantics.

[0030] In this way, the environmental information assigned to each first sub-region reflects both the large-scale climate background and the topography and surface information of its location.

[0031] S2, using a pre-trained thickness prediction model, processes the environmental information of each first sub-region to obtain the predicted thickness of the active layer of each first sub-region.

[0032] It should be understood that when conducting spatial mapping of active layer thickness in permafrost regions, servers typically employ only a single machine learning model, such as random forest or support vector machine. It should be noted that different models have varying responsiveness to complex environmental factors such as topographic relief, vegetation variation, and soil differences. Furthermore, the actual areas being analyzed often contain multiple heterogeneous units, including alpine meadows, desert steppes, and valley transition zones. In such cases, if the server consistently uses a single model, the prediction results may be better in some geomorphic areas but significantly biased in others. Therefore, determining the thickness prediction model solely based on subjective preferences or local accuracy metrics cannot guarantee the overall reliability and spatial stability of the regional mapping results.

[0033] Therefore, the risk diagnosis method for active layer thickness spatial mapping provided in this embodiment also includes a training method for the thickness prediction model. In the training method, the server acquires a sample dataset and divides it into a sample set and a test set; multiple candidate models are trained simultaneously using the sample set; based on the test results of the multiple candidate models in the test set, the model with the best test result is selected as the thickness prediction model.

[0034] This can be understood as the server comparing the actual performance of various machine learning models under the same data and evaluation criteria, thereby selecting the model that is most suitable for the current active layer thickness of the region to be analyzed.

[0035] Specifically, the server can acquire the constructed pixel-level active layer thickness sample dataset and divide it into a sample set and an independent test set according to a fixed ratio. Then, it can use this sample set to train multiple candidate models simultaneously. These candidate models include Random Forest (RF) model, Extra Trees (ET) model, Support Vector Machine (SVM) model, K-Nearest Neighbors (KNN) model, eXtreme Gradient Boosting (XGBoost) model, Light Gradient Boosting Machine (LightGBM) model, and Categorical Boosting (CatBoost) model.

[0036] During this process, the server can perform parameter search for each candidate model and evaluate its stability using a 5-fold × 40 repeated cross-validation method. Model performance can be comprehensively evaluated based on one or more of the following indicators: mean absolute error (MAE), root mean square error (RMSE), residual standard deviation σ, coefficient of determination R², adjusted R², and Willmott Index of Agreement (WIA).

[0037] Through comprehensive comparison, this embodiment found that the CatBoost model achieved an RMSE of 0.267m and an R² of 0.813 on the independent test set. Its prediction results not only demonstrated high overall accuracy but also accurately reproduced the active layer thickness gradient distribution under different vegetation types. For example, the active layer thickness gradient was 1.74±0.46 m for marsh meadow, 2.13±0.48 m for alpine meadow, 2.37±0.56 m for alpine steppe, and 2.82±0.79 m for desert steppe.

[0038] Therefore, in this embodiment, the CatBoost model is ultimately selected as the thickness prediction model.

[0039] Furthermore, the sample dataset includes sample environment information of multiple sample sub-regions in the sample region and the corresponding active layer sample thickness. Therefore, when acquiring the sample dataset, the server can allocate multiple original active layer thicknesses collected from the sample region to multiple second sub-regions according to the collection location; based on the allocation results of multiple original active layer thicknesses, multiple sample sub-regions are selected from the multiple second sub-regions, where each sample sub-region is a second sub-region allocated with original active layer thickness; the active layer sample thickness of each sample sub-region is obtained based on the original active layer thickness allocated to each sample sub-region; and the sample environment information of each sample sub-region is obtained based on the location of each sample sub-region.

[0040] It should be noted that the sample area and the area to be analyzed can come from different regions of the same region, or from the same region. For example, the sample area and the area to be analyzed can both come from different regions of the hot spring area in northeastern Qinghai-Tibet Plateau, or both can be from the same region of the hot spring area in northeastern Qinghai-Tibet Plateau.

[0041] It should be noted that field observation of the active layer thickness in permafrost regions is very difficult. Although traditional point measurement methods such as drilling can measure the active layer thickness with high accuracy, they are costly, inefficient, and the number of samples obtained is extremely limited. While ground-penetrating radar can continuously collect data along the profile, its records are dense trajectory points with irregular spatial locations and uneven density, resulting in limited accuracy.

[0042] Therefore, the multiple original active layer thicknesses include multiple drilling thicknesses and multiple ground-penetrating radar thicknesses. When obtaining the active layer sample thickness of each sample sub-region based on the original active layer thickness of each sample sub-region, the server can fuse the original active layer thicknesses of each sample sub-region to obtain the active layer sample thickness of the sample sub-region.

[0043] In practical applications, multiple drilling thicknesses can be measured by methods such as pit observation, steel rod detection, shallow drilling, and temperature probe observation. Although these methods can obtain relatively reliable observation results at the point scale, they usually have problems such as high field costs, limited spatial coverage, insufficient number of sample points, and weak ability to represent the continuous region.

[0044] Therefore, this embodiment further introduces ground-penetrating radar (GPR) to collect GPR thickness data, complementing the drilling thickness range. Specifically, the server can receive continuous profile observation data from the GPR and verify it with test pit or borehole data. The GPR survey uses a ProEX control unit, the reflection profile uses a 100 MHz shielded antenna to identify the continuous reflection interface corresponding to the bottom boundary of the underground active layer, and the WARR observation uses a 100 MHz unshielded antenna to determine the electromagnetic wave propagation velocity on different vegetation plots. Therefore, for sampling points with suitable conditions, the server can directly calibrate the radar wave velocity based on the measured data from the pit or borehole; for areas lacking measured conditions, the server can assign empirical wave velocity parameters according to vegetation type.

[0045] Based on the corrected wave velocity, the server can calculate the active layer thickness at each point according to the following conversion relationship between Two-Way Travel Time (TWTT) and radar wave velocity:

[0046] In the formula, Indicates the thickness of the active layer. Indicates the speed of radar waves. Indicates round-trip travel time.

[0047] Based on this, the server can divide the entire sample area into multiple second sub-regions, each of which is a 30 m × 30 m grid cell; and select multiple sample sub-regions from the multiple second sub-regions, where each sample sub-region is a second sub-region that contains at least one GPR profile interpretation record or one earth pit measured point.

[0048] Based on the above description of drilling thickness and ground-penetrating radar thickness, this embodiment unifies and structures the original active layer thicknesses from different sources to form sample sub-regions suitable for machine learning modeling.

[0049] In practical applications, the server can statistically analyze the original active layer thicknesses falling within the same grid cell range, calculate their average value under the premise of meeting the minimum sampling density requirement, and use it as the active layer sample thickness of the sample sub-region, with the cell center coordinates as its spatial location; at the same time, the server can also attach attribute information to each sample sub-region, including the number of GPR records in the cell, the representative active layer thickness value, the sample source identifier, and the quality control identifier.

[0050] In this process, the raw active layer thickness processed by the server includes both drilling thickness from point observations such as drilling and ground-penetrating radar (GPR) thickness interpreted from GPR profiles. For example, in this embodiment, from 128 GPR profiles covering different vegetation types and landform units (most of which are over 500 meters long, and some reach 3 kilometers), after depth conversion, pixel aggregation, and quality screening, 810 structurally complete, spatially accurate, and traceable sample sub-regions are finally obtained. Each sample sub-region corresponds to a set of active layer sample thicknesses and sample environmental information of its location.

[0051] In this way, the server merges the original data, which are of different forms, uneven density, and varying quality, to obtain complementary active layer sample thicknesses.

[0052] In the above embodiments, the thickness prediction model and its predicted active layer thickness were introduced. The following will continue with... Figure 1 Step S3 will be explained below: S3 compares the measured thickness of the active layer of each of the multiple reference sub-regions with the predicted thickness of the corresponding active layer to obtain multiple thickness confidence levels.

[0053] Each reference sub-region represents a sub-region with the measured thickness of the active layer.

[0054] In practical applications, the server can extract the known measured thickness of the active layer for each reference sub-region and synchronously call the thickness prediction model to output the predicted thickness of the active layer at that sub-region location. Based on this, a subtraction operation can be performed on each pair of measured and predicted values, that is, the measured thickness is subtracted from the predicted thickness to obtain a single thickness residual, which is used as the thickness confidence level. The thickness residual is a real number with a positive or negative sign. A positive value indicates that the model underestimates the actual thickness, and a negative value indicates that the model overestimates the actual thickness. Its absolute value directly represents the magnitude of the prediction error at that reference sub-region. The larger the value, the lower the confidence level, and vice versa.

[0055] In addition, the server can generate a spatial distribution map of the active layer thickness in the region to be analyzed, based on the measured and predicted active layer thicknesses of multiple reference sub-regions. Users can then visually understand which locations in the region to be analyzed have thicker active layers and which have thinner active layers.

[0056] Based on the above explanation of the thickness reliability, we will continue to discuss... Figure 1 Step S4 will be explained below: S4. Based on the thickness confidence of multiple reference sub-regions, obtain the thickness confidence of the remaining sub-regions in the multiple first sub-regions.

[0057] In this embodiment, the server can use the thickness confidence level at the known reference sub-region as a control point and use spatial interpolation to calculate the estimated prediction deviation value on the first sub-region where no actual measurement points are set.

[0058] In practical applications, the server can first obtain the set of thickness confidence values ​​corresponding to all reference sub-regions. These reference sub-regions are spatially discrete and their number is far less than the total number of all first sub-regions in the region to be analyzed. Therefore, the spatial coordinates of each reference sub-region and its corresponding thickness confidence value are used to form a set of spatial sampling points. Based on this, a mature spatial interpolation algorithm can be used to extrapolate these discrete confidence values ​​to obtain their respective thickness confidence values.

[0059] It should be understood that this thickness confidence level not only reflects the error at known points, but also characterizes the spatial differences in the reliability of model predictions across the entire region.

[0060] The study also found that in related technologies, spatial distribution maps of active layer thickness in permafrost regions are typically output only as a single continuous raster map. The final results generated by the server generally only include the predicted thickness values ​​for each first sub-region, without simultaneously providing information on the spatial reliability of these values. It should be noted that due to uneven sample distribution, abrupt changes in environmental gradients, and inherent limitations of the model, the prediction accuracy varies significantly across different first sub-regions. For example, the prediction results for vegetation boundaries, valley transition zones, or sparsely sampled areas often have large deviations, but these differences are not identified or expressed. In this situation, users cannot intuitively determine which areas' mapping results can be directly used for engineering decisions or ecological assessments, and which areas require cautious reference or even supplementary investigation.

[0061] In view of this, such as Figure 2 As shown, the risk diagnosis method for active layer thickness spatial mapping provided in this embodiment further includes: S5 generates a grid map that corresponds one-to-one with multiple first sub-regions.

[0062] S6. For each first sub-region, select the corresponding target color from the color library based on the thickness confidence level of the first sub-region.

[0063] S7, use the target color to color-mark the target grid in the grid diagram corresponding to the first sub-region.

[0064] S8, after color-coding the grid map, serves as the active layer thickness risk map for the region to be analyzed.

[0065] This can be understood as the server being able to transform the reliability differences in model prediction results into an intuitive spatial visualization, thereby forming an activity layer thickness risk map corresponding to the spatial distribution map of activity layer thickness.

[0066] In practical applications, the server can first generate a mesh map that perfectly matches the region to be analyzed and corresponds one-to-one with multiple first sub-regions. Each target mesh in this mesh map is completely consistent with the first sub-region in terms of spatial location, size, and coordinate system. Then, for each first sub-region, a target color that matches its deviation level can be selected from a preset color library based on its obtained thickness confidence level. For example, mapping a region with lower thickness confidence level to a cool color (e.g., blue) indicates that the prediction result is more reliable; mapping a region with higher thickness confidence level to a warm color (e.g., red or orange) indicates that the first sub-region has a higher risk.

[0067] The server can use the target color to fill the corresponding target grid in the grid diagram, ensuring that the color distribution strictly corresponds to the thickness confidence spatial pattern; the grid diagram with completed color marking is output as an active layer thickness risk map.

[0068] It should be understood that this activity layer thickness risk map can be used to characterize the spatial reliability differences of model prediction results. For example, users can clearly identify high-risk prediction areas such as vegetation boundaries, valley transition zones, sparse sample areas, and abrupt changes in environmental gradients. Furthermore, the activity layer thickness risk map can be used in conjunction with sample distribution density to determine priority areas for supplementary sampling, and supplementary observations can be carried out in high-risk and sparse sample areas. The newly added data can then be remodeled, achieving a closed-loop optimization of "risk identification, supplementary sampling, and updated mapping".

[0069] Based on the same inventive concept as the risk diagnosis method for active layer thickness spatial mapping provided in this embodiment, this embodiment also provides a risk diagnosis device for active layer thickness spatial mapping. This device includes at least one software functional module that can be stored in a memory or embedded in an electronic device. A processor in the electronic device executes the executable module stored in the memory. For example, the software functional module and computer program included in this device. Please refer to... Figure 3 Functionally, the device may include: Environmental information module 11 is used to acquire environmental information of each of the multiple first sub-regions in the area to be analyzed; The thickness prediction module 12 is used to process the environmental information of each first sub-region using a pre-trained thickness prediction model to obtain the predicted thickness of the active layer of each first sub-region. The risk prediction module 13 is used to compare the measured thickness of the active layer of each of the multiple reference sub-regions with the corresponding predicted thickness of the active layer to obtain the thickness confidence of the multiple reference sub-regions, wherein each reference sub-region represents a sub-region with the measured thickness of the active layer. The risk prediction module 13 is also used to obtain the thickness confidence of the remaining sub-regions in the multiple first sub-regions based on the thickness confidence of multiple reference sub-regions.

[0070] In this embodiment, the environmental information module 11 is used to implement... Figure 1 In step S1, the thickness prediction module 12 is used to implement Figure 1 In step S2, the risk prediction module 13 is used to implement... Figure 1 Steps S3 and S4 in the above process. Therefore, for a detailed description of each of the above modules, please refer to the specific implementation methods of the corresponding steps.

[0071] Optionally, the risk prediction module 13 obtains the thickness confidence of the remaining sub-regions in the multiple first sub-regions based on the thickness confidence of multiple sub-regions, including: Based on the thickness confidence of multiple sub-regions, interpolation is performed on the remaining sub-regions in the multiple first sub-regions to obtain the thickness confidence of the remaining sub-regions in the multiple first sub-regions.

[0072] Optionally, the risk prediction module 13 is also used for: Generate a grid diagram that corresponds one-to-one with multiple first sub-regions; For each first sub-region, select the corresponding target color from the color library based on the thickness confidence level of the first sub-region; Use the target color to color-mark the target grid in the grid diagram that corresponds to the first sub-region; After color-coding the grid diagram, it serves as the active layer thickness risk map for the region to be analyzed.

[0073] Optionally, the risk diagnosis device for active layer thickness spatial mapping includes a model training module for a thickness prediction model. The model training module is used to acquire a sample dataset and divide it into a sample set and a test set; to train multiple candidate models simultaneously using the sample set; and to select the model with the best test results from the multiple candidate models in the test set as the thickness prediction model.

[0074] Optionally, the sample dataset includes sample environment information of multiple sample sub-regions within the local area and the corresponding active layer sample thickness. The model training module obtains the sample dataset in the following ways: The sample region is divided into multiple second sub-regions; The original active layer thicknesses collected from the sample area are distributed to multiple second sub-regions according to the collection location; Based on the allocation results of multiple original active layer thicknesses, multiple sample sub-regions are selected from multiple second sub-regions, where each sample sub-region is a second sub-region that has been allocated an original active layer thickness. The active layer sample thickness of each sample sub-region is obtained based on the original active layer thickness of each sample sub-region. Based on the location of each sample sub-region, obtain the sample environment information of each sample sub-region.

[0075] Optionally, the multiple original active layer thicknesses include multiple drilling thicknesses and multiple ground-penetrating radar thicknesses. The model training module obtains the active layer sample thickness for each sample sub-region based on the original active layer thicknesses allocated to each sample sub-region, including the following methods: For each sample sub-region, the original active layer thicknesses of the sample sub-region are fused to obtain the active layer sample thickness of the sample sub-region.

[0076] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0077] It should also be understood that if the above embodiments are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0078] Therefore, this embodiment also provides a storage medium, which is a computer-readable storage medium. This storage medium stores a computer program, which, when executed by a processor, implements the risk diagnosis method for active layer thickness spatial mapping provided in this embodiment. The storage medium can be any medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0079] Please refer to Figure 4The electronic device provided in this embodiment may include a processor 22 and a memory 21. The memory 21 stores a computer program, and the processor reads and executes the computer program in the memory 21 corresponding to the above-described embodiments to implement the risk diagnosis method for active layer thickness spatial mapping provided in this embodiment.

[0080] See also Figure 4 The electronic device also includes a communication unit 23. The memory 21, processor 22 and communication unit 23 are electrically connected to each other directly or indirectly through system bus 24 to realize data transmission or interaction.

[0081] The memory 21 can be an information recording device based on any electronic, magnetic, optical, or other physical principles, used to record execution instructions, data, etc. In some embodiments, the memory 21 can be, but is not limited to, volatile memory, non-volatile memory, memory drive, etc.

[0082] In some embodiments, the volatile memory may be random access memory (RAM); in some embodiments, the non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory, etc.; in some embodiments, the storage drive may be a disk drive, solid-state drive, any type of storage disk (such as optical disc, DVD, etc.), or similar storage media, or a combination thereof.

[0083] The communication unit 23 is used to send and receive data over a network. In some embodiments, the network may include a wired network, a wireless network, a fiber optic network, a telecommunications network, an intranet, the Internet, a local area network (LAN), a wide area network (WAN), a wireless local area network (WLAN), a metropolitan area network (MAN), a public switched telephone network (PSTN), a Bluetooth network, a ZigBee network, or a near field communication (NFC) network, or any combination thereof. In some embodiments, the network may include one or more network access points. For example, the network may include wired or wireless network access points, such as base stations and / or network switching nodes, through which one or more components of the service request processing system can connect to the network to exchange data and / or information.

[0084] The processor 22 may be an integrated circuit chip with signal processing capabilities, and may include one or more processing cores (e.g., a single-core processor or a multi-core processor). By way of example only, the processor described above may include a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), an Application Specific Instruction-set Processor (ASIP), a Graphics Processing Unit (GPU), a Physics Processing Unit (PPU), a Digital Signal Processor (DSP), a Field Programmable Gate Array (FPGA), a Programmable Logic Device (PLD), a controller, a microcontroller unit, a Reduced Instruction Set Computing (RISC) computer, or a microprocessor, or any combination thereof.

[0085] Understandable. Figure 4The structure shown is for illustrative purposes only. Electronic devices may also have more advanced features. Figure 4 Showing more or fewer components, or having with Figure 4 The different configurations shown. Figure 4 The components shown can be implemented using hardware, software, or a combination thereof.

[0086] It should be understood that the apparatus and methods disclosed in the above embodiments can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0087] The above descriptions are merely various embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A risk diagnosis method for spatial mapping of active layer thickness, characterized in that, The method includes: Obtain the environmental information of each of the multiple first sub-regions in the region to be analyzed; The environmental information of each first sub-region is processed using a pre-trained thickness prediction model to obtain the predicted active layer thickness of each first sub-region. The measured active layer thickness of each of the multiple reference sub-regions is compared with the corresponding predicted active layer thickness to obtain the thickness confidence of the multiple reference sub-regions, wherein each of the reference sub-regions represents a sub-region with a measured active layer thickness; Based on the thickness confidence of the plurality of reference sub-regions, the thickness confidence of the remaining sub-regions in the plurality of first sub-regions is obtained.

2. The risk diagnosis method for spatial mapping of active layer thickness according to claim 1, characterized in that, Based on the thickness confidence levels of the plurality of reference sub-regions, the thickness confidence levels of the remaining sub-regions within the plurality of first sub-regions are obtained, including: Based on the thickness confidence level of the plurality of reference sub-regions, interpolation is performed on the remaining sub-regions in the plurality of first sub-regions to obtain the thickness confidence level of the remaining sub-regions in the plurality of first sub-regions.

3. The risk diagnosis method for spatial mapping of active layer thickness according to any one of claims 1-2, characterized in that, The method further includes: Generate a grid diagram that corresponds one-to-one with the plurality of first sub-regions; For each of the first sub-regions, a corresponding target color is selected from the color library based on the thickness confidence level of the first sub-region; The target color is used to color-mark the target grid in the grid diagram that corresponds to the first sub-region; The grid map, after being color-coded, serves as the active layer thickness risk map for the region to be analyzed.

4. The risk diagnosis method for spatial mapping of active layer thickness according to claim 1, characterized in that, The method also includes a training method for the thickness prediction model, the training method comprising: Obtain the sample dataset and divide it into a sample set and a test set; Multiple candidate models can be trained simultaneously using the sample set; Based on the test results of multiple candidate models in the test set, the model with the best test results is selected as the thickness prediction model.

5. The confidence method for spatial mapping of active layer thickness according to claim 4, characterized in that, The candidate models are Random Forest, Extreme Random Tree, Support Vector Machine, K-Nearest Neighbors, Extreme Gradient Boosting XGBoost, Lightweight Gradient Boosting Machine (LightGBM), and Classification-Enhanced Gradient Boosting CatBoost.

6. The risk diagnosis method for spatial mapping of active layer thickness according to claim 4, characterized in that, The sample dataset includes sample environment information of multiple sample sub-regions within the sample region and the corresponding active layer sample thickness. Obtaining the sample dataset includes: The sample region is divided into multiple second sub-regions; The original active layer thicknesses collected from the sample region are distributed to the multiple second sub-regions according to the collection location; Based on the allocation results of the multiple original active layer thicknesses, multiple sample sub-regions are selected from the multiple second sub-regions, wherein each sample sub-region is a second sub-region that has been allocated the original active layer thickness; The active layer sample thickness of each sample sub-region is obtained based on the original active layer thickness of each sample sub-region. Based on the location of each sample sub-region, obtain the sample environment information of each sample sub-region.

7. The risk diagnosis method for spatial mapping of active layer thickness according to claim 6, characterized in that, The multiple original active layer thicknesses include multiple drilling thicknesses and multiple ground-penetrating radar thicknesses. The active layer sample thickness for each sample sub-region is obtained based on the original active layer thicknesses allocated to each sample sub-region, including: For each sample sub-region, the original active layer thicknesses of the sample sub-region are fused to obtain the active layer sample thickness of the sample sub-region.

8. A risk diagnosis device for spatial mapping of active layer thickness, characterized in that, The device includes: The environmental information module is used to obtain the environmental information of each of the multiple first sub-regions in the area to be analyzed; The thickness prediction module is used to process the environmental information of each of the first sub-regions using a pre-trained thickness prediction model to obtain the predicted active layer thickness of each of the first sub-regions. The risk prediction module is used to compare the measured thickness of the active layer of each of the multiple reference sub-regions with the corresponding predicted thickness of the active layer to obtain the thickness confidence of the multiple reference sub-regions, wherein each of the reference sub-regions represents a sub-region with a measured thickness of the active layer; The risk prediction module is further configured to obtain the thickness confidence of the remaining sub-regions in the plurality of first sub-regions based on the thickness confidence of the plurality of reference sub-regions.

9. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the risk diagnosis method for spatial mapping of active layer thickness as described in any one of claims 1-7.

10. An electronic device, characterized in that, The electronic device includes a processor and a memory, the memory storing a computer program, which, when executed by the processor, implements the risk diagnosis method for spatial mapping of active layer thickness as described in any one of claims 1-7.

Images