Session rainfall water and soil loss prediction method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of water and soil conservation, and discloses a session rainfall water and soil loss prediction method and device, electronic equipment and a storage medium, and the method comprises the steps: setting classification rules for runoff plot site conditions according to different slope grades, vegetation coverage and engineering measure types, historical session rainfall data and water and soil loss monitoring data of runoff plots with different site condition types are obtained, and a multi-factor combination data set is constructed; based on the multi-factor combination data set, training a preset deep learning model to obtain a session rainfall water and soil loss prediction model; according to the method, a target area is divided into natural land parcels with different site conditions on the basis of a classification rule, and the model is driven to carry out session rainfall water and soil loss prediction of each natural land parcel on the basis of the collected rainfall forecast data, so that a dynamic mechanism of a rainfall-runoff-erosion process is effectively captured; the limitation that a traditional empirical model is difficult to predict single rainfall water and soil loss is overcome, and the refinement degree of prediction is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of soil and water conservation technology, specifically to methods, devices, electronic equipment, and storage media for predicting soil erosion during rainfall events. Background Technology

[0002] Soil erosion is a severe ecological problem facing the world, seriously restricting regional ecological security and sustainable development. Accurate prediction is key to effective prevention and control. Current soil erosion analysis and calculations largely rely on the Universal Soil Loss Equation (USLE) and its derivative models (such as RUSLE (Revised Universal Soil Loss Equation) and CSLE (Chinese Soil Loss Equation)). These models are based on long-term statistical patterns and are suitable for estimating erosion at the annual average scale, but they are difficult to apply to the calculation and prediction of dynamic processes of rapid soil erosion triggered by a single rainfall event. In recent years, extreme weather events have become more frequent, and existing research on soil erosion caused by single rainfall events has certain shortcomings in analysis and prediction. Summary of the Invention

[0003] This invention provides a method, apparatus, electronic device, and storage medium for predicting soil erosion under single-rainfall conditions, in order to solve the problem of insufficient soil erosion prediction capability under single-rainfall conditions in the prior art.

[0004] In a first aspect, the present invention provides a method for predicting soil erosion during a single rainfall event, the method comprising: Classification rules are set for the site conditions of runoff plots according to different slope grades, vegetation coverage, and types of engineering measures; Historical rainfall data and soil erosion monitoring data of runoff plots with different site conditions in the target area or similar topographic and geomorphological areas are obtained, and a multi-factor combination dataset is constructed based on the historical rainfall data and soil erosion monitoring data. Based on a multi-factor combination dataset, a pre-set deep learning model is trained to obtain a precipitation-based soil erosion prediction model. Based on classification rules, the target area is divided into natural plots with different site conditions. Based on the collected forecast rainfall data, the event-based rainfall soil erosion prediction model is driven to predict the event-based rainfall soil erosion of each natural plot.

[0005] This invention provides a method for predicting soil erosion under inter-rainfall conditions. By integrating multi-factor site condition classification with a deep learning model, it effectively captures the dynamic mechanism of the rainfall-runoff-erosion process, overcoming the shortcomings of traditional empirical models in predicting soil erosion under inter-rainfall conditions. This method reliably extends the prediction capabilities from annual-scale monitoring to inter-rainfall-scale prediction. Furthermore, it innovatively proposes a natural land parcel division technique and supports rainfall forecast input, significantly improving the precision, spatiotemporal adaptability, and operational capability of the prediction. This provides a scientific and efficient technical tool for soil erosion control decision-making and solves the problem of insufficient soil erosion prediction capability under inter-rainfall conditions in existing technologies.

[0006] In one optional implementation, historical rainfall data and soil erosion monitoring data of runoff plots with different site conditions in the target area or similar topographic regions are acquired, including: Based on the classification rules for site conditions of runoff plots according to different slope grades, vegetation cover and engineering measures, historical rainfall data and soil erosion monitoring data are classified according to site condition types.

[0007] The present invention provides a method for predicting soil erosion during rainfall events. Through a preset multi-factor classification rule, historical rainfall monitoring data are systematically organized and classified according to different site conditions (combinations of slope, vegetation, and engineering measures). This transforms the raw data into a training sample set with a clear structure, well-defined categories, and ease of model learning, laying an accurate data foundation for the subsequent construction of targeted prediction models.

[0008] In one alternative implementation, historical rainfall data includes rainfall characteristics, and soil erosion monitoring data includes slope runoff and soil erosion. Based on historical rainfall data and soil erosion monitoring data, a multi-factor combined dataset was constructed, including: Using the characteristics of each rainfall event as input data and slope runoff and soil erosion as output data, and constructing input-output data pairs based on the site condition types of the input and output data, the input-output data pairs bound to the site condition types are used as a multi-factor combined dataset.

[0009] This invention provides a method for predicting soil erosion based on rainfall events. This method uses structured inputs that bind rainfall event characteristics with corresponding slope runoff and soil erosion based on site conditions. The output data pairs not only establish a direct mapping relationship from rainfall-driven soil erosion response, but also realize differentiated expressions of this mapping relationship under different site conditions, thereby enabling subsequent models to learn more targeted and reliable prediction patterns and effectively improving the prediction accuracy of the models.

[0010] In one optional implementation, a pre-defined deep learning model is trained based on a multi-factor combination dataset to obtain a soil erosion prediction model, including: The multi-factor combination dataset is input into the preset deep learning model, which captures the features of rainfall-runoff-erosion through sequence learning and establishes a nonlinear mapping relationship between the features of each rainfall event and the slope runoff and soil erosion under different site conditions, thus obtaining a rainfall-soil loss prediction model.

[0011] This invention provides a method for predicting soil erosion during rainfall events. By employing a deep learning model with sequence learning capabilities and scientifically dividing the training, validation, and test sets, the model can effectively learn the dynamic temporal patterns of the rainfall-runoff-erosion process and establish accurate nonlinear mappings for different site conditions. This significantly improves the model's ability to simulate and predict soil erosion during rainfall events, overcoming the shortcomings of traditional static statistical models in temporal dynamic prediction.

[0012] In one alternative implementation, based on classification rules, the target area is divided into natural plots with different site conditions, including: Obtain land use data for the target area, and based on the land use data, remove hardened surface areas such as villages and roads to obtain the natural surface range to be delineated; Acquire digital elevation data, vegetation data, and soil and water conservation engineering measures data of the target area, and generate multi-source spatial data with different slope levels, vegetation cover levels, and soil and water conservation engineering measures types corresponding to the classification rules based on the digital elevation data, vegetation data, and soil and water conservation engineering measures data. By unifying multi-source spatial data into the same geographic coordinate system and spatial resolution, natural land parcels with different site conditions can be obtained.

[0013] This invention provides a method for predicting soil erosion based on rainfall events. By systematically integrating and refining multi-source spatial data, and after eliminating interference from hardened surfaces, it integrates information on topography, vegetation, and engineering measures to achieve scientific and automated plot division of the natural surface within the prediction area. This deconstructs large, complex areas into multiple prediction units with uniform internal site conditions. This not only ensures the spatial relevance and scale applicability of subsequent predictions but also provides a reliable spatial carrier for soil erosion models to move from point-scale monitoring to regional-scale simulation.

[0014] In one optional implementation, based on collected forecast rainfall data, a soil erosion prediction model is driven to predict soil erosion during rainfall events for each natural plot, including: Using forecasted rainfall data, different slope grades, vegetation cover grades, and types of soil and water conservation engineering measures as inputs, the soil and water loss prediction model is driven to make batch predictions of slope runoff and soil erosion in each natural plot under the conditions of each rainfall event, so as to obtain the soil and water loss prediction results for each natural plot in the target area.

[0015] The present invention provides a method for predicting soil erosion during a single rainfall event. By synchronously inputting forecast rainfall data and high-resolution spatial attribute data, it can automatically drive parallel calculations of dedicated models for different natural plots. While ensuring that the prediction results strictly match the actual conditions of each plot of land, it can complete batch predictions for the entire region at once, significantly improving the timeliness, accuracy and operational capability of the prediction.

[0016] In one alternative implementation, the forecast rainfall data supports the following input methods: Read uniform rainfall patterns from hourly rainfall sequences in text format or spatially distributed rainfall patterns from rainfall raster sequences in GeoTIFF format.

[0017] The present invention provides a method for predicting soil erosion based on rainfall events. This method supports both simplified and rapid input of uniform rainfall patterns and refined driving of spatially distributed rainfall. This allows the prediction method to flexibly cope with various actual meteorological conditions, such as uniform regional rainfall and uneven spatiotemporal distribution, and significantly enhances the applicability, operability, and reliability of the prediction results.

[0018] Secondly, the present invention provides a device for predicting soil erosion during a rainfall event, the device comprising: The multi-factor combined dataset construction module is used to set classification rules for runoff plot site conditions according to different slope levels, vegetation cover and engineering measure types, obtain historical rainfall data and soil erosion monitoring data of runoff plots with different site condition types in the target area or similar topographic and geomorphological areas, and construct multi-factor combined datasets based on historical rainfall data and soil erosion monitoring data. The model training module is used to train a preset deep learning model based on a multi-factor combination dataset to obtain a rainfall-based soil erosion prediction model. The soil and water loss prediction module is used to divide the target area into natural plots with different site conditions according to classification rules, and drive the precipitation soil and water loss prediction model to predict precipitation soil and water loss for each natural plot based on the collected forecast rainfall data.

[0019] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the above-described method for predicting soil erosion during rainfall or any corresponding embodiment.

[0020] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the method for predicting soil erosion during rainfall as described in the first aspect or any corresponding embodiment.

[0021] Fifthly, the present invention provides a computer program product, including computer instructions for causing a computer to execute the method for predicting soil erosion during rainfall as described in the first aspect or any corresponding embodiment. Attached Figure Description

[0022] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of an application scenario according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the first process of a rainfall-based soil erosion prediction method according to an embodiment of the present invention. Figure 3 This is a schematic diagram of the second process of the method for predicting soil erosion by rainfall according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the third process of the method for predicting soil erosion by rainfall according to an embodiment of the present invention. Figure 5 This is a schematic diagram of the slope runoff prediction grid results in the rainfall-soil erosion prediction method according to an embodiment of the present invention. Figure 6 This is a schematic diagram of the soil erosion prediction grid results in the rainfall-soil loss prediction method according to an embodiment of the present invention. Figure 7 This is a structural block diagram of a rainfall-soil erosion prediction device according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0026] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0027] As an optional application scenario of this invention, such as Figure 1 As shown, application 101 is installed in terminal device 110, and user 130 can interact with application 101 through terminal device 110 and / or access device of terminal device 110.

[0028] For example, application 101 can be any application that provides question-and-answer related services. For instance, application 101 could be a question-and-answer interactive application, such as a text-to-text application, an image-to-text application, etc. Figure 1 In the application scenario shown, if application 101 is active, the terminal device 110 can display the interface 102 of application 101. The interface 102 may include various pages that application 101 can provide, such as interactive pages, settings pages, query pages, etc.

[0029] In some embodiments, terminal device 110 is communicatively connected to server 120 to provide services to application 101. Terminal device 110 may be a mobile terminal, fixed terminal, or portable terminal, etc., including but not limited to mobile phones, desktop computers, laptop computers, multimedia tablets, e-book devices, gaming devices, or any combination thereof, including accessories and peripherals of these devices or any combination thereof. In some embodiments, terminal device 110 may also support any type of interface, and server 120 may be various types of computing systems or servers capable of providing computing power, including but not limited to mainframes, edge computing nodes, computing devices in cloud environments, etc.

[0030] It should be noted that, Figure 1 This is merely an example of an application scenario and does not limit the scope of protection of this invention.

[0031] The embodiments of the present invention will now be described with reference to the accompanying drawings. It should be understood that the pages shown in the drawings are merely examples, and various page designs are possible in practice. The various graphic elements on the page may have different arrangements and different visual representations; one or more elements may be omitted or replaced, and one or more other elements may also be present, without any limitation in the embodiments of the present invention. Furthermore, the embodiments described below primarily pertain to terminal device 110. It should be understood that the actions described relative to terminal device 110 can be performed by application 101 on terminal device 110, or can be performed by application 101 in conjunction with its server (e.g., server 120).

[0032] The mechanisms of soil erosion are complex, spanning multiple disciplines such as soil science, hydrology, and vegetation, making the construction and solution of accurate physical mechanism models extremely challenging. Deep learning, as a powerful machine learning technology, has opened up new avenues for modeling complex environmental processes with its excellent nonlinear mapping and high-efficiency computing power, and has achieved remarkable results in hydrological and meteorological forecasting. However, its application in the field of soil erosion prediction is still in its early stages.

[0033] This invention provides a method for predicting soil erosion during a single rainfall event. By integrating multi-factor site condition classification with a deep learning-based neural network model, it effectively captures the dynamic mechanism of the rainfall-runoff-erosion process, overcomes the limitations of traditional empirical models in predicting soil erosion during single rainfall events, improves the simulation and forecasting capabilities for soil erosion under single rainfall conditions, and solves the technical problem of insufficient prediction capabilities for soil erosion under single rainfall conditions.

[0034] According to an embodiment of the present invention, a method for predicting soil erosion during a rainfall event is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0035] This embodiment provides a method for predicting soil erosion based on rainfall events, which can be used in the aforementioned electronic or terminal devices. Figure 2 This is a flowchart of a rainfall-based soil erosion prediction method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Set classification rules for the site conditions of runoff plots according to different slope levels, vegetation coverage, and types of engineering measures.

[0036] This step establishes a refined site condition classification rule system based on multi-factor combination: that is, by pre-setting specific classification standards for three key environmental factors: slope level (e.g., 5°~15°, 15°~20°, etc.), vegetation cover level (e.g., below 30%, 30%~60%, etc.), and soil and water conservation engineering measure type (e.g., terraces, horizontal steps, etc.), and systematically combining these three types of factors to form a complete and clear classification rule; then, based on this rule, the site attributes of all runoff plots are identified and classified to obtain different site condition types, thus laying the data foundation for the subsequent construction of multi-factor datasets.

[0037] Step S202: Obtain historical rainfall data and soil erosion monitoring data of runoff plots with different site conditions in the target area or similar topographic and geomorphological areas, and construct a multi-factor combination dataset based on the historical rainfall data and soil erosion monitoring data.

[0038] Specifically, the target area refers to the spatial range within which soil erosion prediction needs to be performed. It is typically a natural geographical unit or administrative unit with clear geographical boundaries that may be affected by rainfall events and thus subject to soil erosion. In other words, the target area is the complete geographical carrier in this method, from data preparation and land parcel delineation to prediction result output, and is the specific spatial object on which the prediction task takes place.

[0039] Similar topography and geomorphology areas refer to areas that have a high degree of consistency with the target area in terms of macroscopic natural geographical environment, thereby ensuring that the dominant processes and mechanisms of soil and water loss are similar.

[0040] Different site condition types refer to the categories obtained by systematically dividing land units according to key environmental factors affecting soil and water loss. Key environmental factors include: slope grade, vegetation cover grade, and type of soil and water conservation engineering measures.

[0041] Historical rainfall data and soil erosion monitoring data refer to paired observation datasets collected over a long period from runoff plots (standardized slope observation units) for a single rainfall event, including: Historical rainfall data mainly includes rainfall (total precipitation in a rainfall event) and rainfall intensity (rainfall intensity, usually hourly time-series data), which represents an independent rainfall event rather than a long-term climate mean.

[0042] Soil and water loss monitoring data refers to the actual soil and water loss results observed in runoff plots corresponding to the above-mentioned rainfall events during the same rainfall event. It mainly includes: Slope runoff: the total volume of surface runoff generated by the rainfall event (unit: cubic meters); Soil erosion: the total amount of soil washed away and carried away by the runoff during the rainfall event (unit: tons).

[0043] They are usually collected and compiled from monitoring databases, research reports, or publicly published literature of relevant units, such as through soil and water conservation monitoring networks.

[0044] This step still follows the classification rules set in step S201 for site conditions of runoff plots based on different slope grades, vegetation cover, and engineering measure types. The multi-factor combination dataset refers to a dataset obtained from the three key environmental factors that determine soil erosion: different slope grades, vegetation cover, and engineering measure types.

[0045] The basic unit of the dataset is the input-output data pair, and each data pair is bound to a key label, namely the site condition type of the runoff plot from which the data originates.

[0046] Step S203: Based on the multi-factor combination dataset, train the preset deep learning model to obtain the rainfall-soil erosion prediction model.

[0047] Specifically, this involves categorizing rainfall events by multiple factors (slope, vegetation, engineering measures). Soil and water loss data are used as training samples to input into a deep learning model, which captures rainfall data through sequence learning. run-off The dynamic temporal characteristics of the erosion process were studied, and a strategy of set training, optimization, and validation was adopted to finally establish a prediction model for different site conditions that can accurately map the characteristics of rainfall events to slope runoff and soil erosion.

[0048] Step S204: Based on classification rules, the target area is divided into natural plots with different site conditions, and based on the collected forecast rainfall data, the precipitation-based soil erosion prediction model is driven to predict precipitation-based soil erosion for each natural plot.

[0049] Specifically, dividing the target area into natural plots with different site conditions means, based on the aforementioned unified classification rules, cutting a complete geographical area (target area) into multiple smaller spatial units with relatively uniform internal site conditions (slope, vegetation, engineering measures), i.e., natural plots.

[0050] Forecasted rainfall data refers to rainfall data that serves as model-driven input and is used to predict soil erosion caused by a possible future rainfall event.

[0051] Driving a soil and water loss prediction model to predict soil and water loss in each natural plot during a rainfall event refers to the process of automatically calling and running a pre-trained soil and water loss prediction model, taking prepared forecast rainfall data and spatial plot data as input, calculating the slope runoff and soil erosion that may occur in each natural plot in the target area during a future rainfall event, and summarizing the overall prediction results for the region.

[0052] The method for predicting soil erosion based on rainfall events provided in this embodiment effectively captures the dynamic mechanism of the rainfall-runoff-erosion process by integrating multi-factor site condition classification with a deep learning model. This overcomes the limitation of traditional empirical models in predicting soil erosion based on rainfall events. Furthermore, it innovatively proposes a natural land parcel division technique, enabling a reliable extension from annual-scale monitoring to rainfall event-scale prediction. It also supports rainfall forecast input, significantly improving the precision, spatiotemporal adaptability, and operational capability of the predictions. This provides a scientific and efficient technical tool for soil erosion control decision-making and solves the technical challenge of insufficient soil erosion prediction capability under rainfall event conditions.

[0053] This embodiment provides a method for predicting soil erosion based on rainfall events, which can be used in the aforementioned electronic or terminal devices. Figure 3 This is a flowchart of a rainfall-based soil erosion prediction method according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: Step S301: Set classification rules for the site conditions of the runoff plots in the target area according to different slope levels, vegetation cover levels and types of soil and water conservation engineering measures.

[0054] Specifically, slope grades, such as 5°~15°, 15°~20°, 20°~25°, and above 25°; Vegetation cover levels, such as below 30%, 30%~60%, 60%~80%, and above 80%; Types of soil and water conservation engineering measures, such as earthen embankment horizontal terraces, stone embankment horizontal terraces, sloping terraces, and horizontal steps.

[0055] The site conditions of the runoff plots in the target area were classified according to the different slope levels, vegetation cover levels, and types of soil and water conservation engineering measures, resulting in different site condition types.

[0056] Step S302: Obtain historical rainfall data and soil erosion monitoring data of runoff plots with different site conditions in the target area or similar topographic and geomorphological areas, and construct a multi-factor combination dataset based on the historical rainfall data and soil erosion monitoring data.

[0057] Specifically, historical rainfall data includes rainfall characteristics, and soil erosion monitoring data includes slope runoff and soil erosion; step S302 above includes: Step S3021: Based on the rules for classifying site conditions of runoff plots according to different slope grades, vegetation cover and engineering measure types, historical rainfall data and soil erosion monitoring data are classified according to site condition types.

[0058] Specifically, according to the above-mentioned preset classification rules: slope grade (5°~15°, 15°~20°, 20°~25°, above 25°), vegetation coverage grade (below 30%, 30%~60%, 60%~80%, above 80%), and soil and water conservation engineering measures type (earthen embankment horizontal terraces, stone embankment horizontal terraces, sloping terraces, horizontal steps, etc.), the historical rainfall and soil erosion monitoring data of slope runoff observation plots in the watershed over many years are collected and organized.

[0059] Step S3022: Using the characteristics of each rainfall event as input data and the slope runoff and soil erosion as output data, and constructing input-output data pairs based on the site condition types to which the input and output data belong, the input-output data pairs bound to the site condition types are used as a multi-factor combined dataset.

[0060] Specifically, the characteristics of each rainfall event include the timing characteristics of the rainfall.

[0061] The model input data uses the temporal characteristics of a single rainfall event, while the corresponding slope runoff and soil erosion are used as the model output target data. Each input-output data set is clearly labeled with the slope, vegetation cover, and engineering measure combination type of the source runoff plot. The rainfall characteristics and soil erosion data labeled with these site types are combined to form an input-output data pair with clear environmental labels. All data pairs categorized and constructed according to site condition type are summarized to form a multi-factor combination dataset for model training.

[0062] Step S303: Based on the multi-factor combination dataset, train the preset deep learning model to obtain a rainfall-based soil erosion prediction model. For details, please refer to [link to details]. Figure 2 Step S203 of the illustrated embodiment will not be described again here.

[0063] Step S304: Based on classification rules, the target area is divided into natural plots with different site conditions. Then, based on the collected forecast rainfall data, the soil erosion prediction model is driven to predict soil erosion during specific rainfall events for each natural plot. For details, please refer to [link to details]. Figure 2 Step S204 of the illustrated embodiment will not be described again here.

[0064] The rainfall-based soil erosion prediction method provided in this embodiment establishes refined classification rules based on a combination of multiple factors (slope, vegetation cover, and engineering measures). This ensures that data organization, model training, and spatial prediction strictly correspond to real site conditions, fundamentally guaranteeing the professionalism of the prediction model and the accuracy of the prediction results. It represents a leap in soil erosion simulation from general empirical estimation to precise categorization. By binding rainfall characteristics with corresponding slope runoff and soil erosion according to site condition types as structured inputs, this method achieves a significant improvement. The output data pairs establish a direct mapping relationship from rainfall-driven soil erosion response.

[0065] This embodiment provides a method for predicting soil erosion based on rainfall events, which can be used in the aforementioned electronic or terminal devices. Figure 4 This is a flowchart of a rainfall-based soil erosion prediction method according to an embodiment of the present invention, such as... Figure 4 As shown, the process includes the following steps: Step S401: Classify the site conditions of runoff plots according to different slope grades, vegetation cover, and types of engineering measures. For details, please refer to [link to relevant documentation]. Figure 3 Step S301 of the illustrated embodiment will not be described again here.

[0066] Step S402: Obtain historical rainfall data and soil erosion monitoring data for runoff plots of different site conditions within the target area or similar topographic regions. Based on the historical rainfall data and soil erosion monitoring data, construct a multi-factor combined dataset. For details, please refer to [link to relevant documentation]. Figure 3 Step S302 of the illustrated embodiment will not be described again here.

[0067] Step S403: Based on the multi-factor combination dataset, train the preset deep learning model to obtain the rainfall-soil erosion prediction model.

[0068] Specifically, this involves categorizing rainfall events by multiple factors (slope, vegetation, engineering measures). Soil and water loss data are used as training samples to input into a deep learning model, which captures rainfall data through sequence learning. run-off The dynamic temporal characteristics of the erosion process were studied, and a strategy of set training, optimization and validation was adopted to finally establish a set rainfall-based soil and water loss prediction model that can accurately map the characteristics of each rainfall event to slope runoff and soil erosion for different site conditions.

[0069] Step S404: Based on preset classification rules, the target area is divided into natural plots with different site conditions, and based on the collected forecast rainfall data, the soil and water loss prediction model is driven to predict the soil and water loss of each natural plot during rainfall events.

[0070] Specifically, step S404 includes: Step S4041: Obtain land use data for the target area, and based on the land use data, remove hardened surface areas of villages and roads to obtain the natural surface range to be delineated.

[0071] Specifically, based on land use data, villages, roads, and other artificially hardened areas are first removed from the predicted area, and the area to be delineated is extracted, which only contains the original natural surface that may be subject to soil erosion.

[0072] Step S4042: Obtain digital elevation data, vegetation data, and soil and water conservation engineering measures data of the target area, and generate multi-source spatial data with different slope levels, vegetation cover levels, and soil and water conservation engineering measures types corresponding to the classification rules based on the digital elevation data, vegetation data, and soil and water conservation engineering measures data.

[0073] Specifically, the Digital Elevation Model (DEM) of the target area is resampled, the slope is calculated, and divided into four slope levels. The Normalized Difference Vegetation Index (NDVI) data of the target area is resampled, converted into vegetation cover, and divided into four vegetation cover levels. The soil and water conservation engineering measures data of the target area are identified and classified according to the type of engineering measures, and converted into spatial data layers that conform to pre-established site classification rules.

[0074] Step S4043: Unify the multi-source spatial data into the same geographic coordinate system and spatial resolution to obtain natural plots with different site conditions.

[0075] Specifically, all spatial data layers are unified to the same geographic coordinate system and spatial resolution. Through overlay analysis and aggregation, multiple natural land parcel units with uniform internal site conditions and complete matching with classification rules are formed.

[0076] Step S4044: The forecasted rainfall data, different slope grades, vegetation cover grades and types of soil and water conservation engineering measures are used as inputs to drive the soil and water loss prediction model to perform batch predictions of slope runoff and soil erosion of each natural plot under the conditions of each rainfall event, so as to obtain the soil and water loss prediction results of each natural plot in the target area.

[0077] The forecast rainfall data supports the following input methods: The forecast rainfall data is analyzed in two ways: first, a uniform rainfall pattern, which reads hourly rainfall sequences in text format; and second, a spatially distributed rainfall pattern, which reads rainfall raster sequences in GeoTIFF format (a TIFF image file format containing geospatial reference information). Projection transformation and resampling are performed on the forecast rainfall data to ensure consistency with the underlying spatial data.

[0078] A batch prediction method is adopted, using forecasted rainfall data and preprocessed slope grade, vegetation cover grade, and soil and water conservation engineering measure classification as input. Based on the attribute characteristics of each pixel, the corresponding deep learning model is automatically selected for prediction, calculating the slope runoff and soil erosion of all pixels. Spatially summing these values ​​yields the watershed / regional slope runoff and soil erosion, generating a spatial distribution map. In other words, by automatically matching the site conditions of each pixel (plot) with the corresponding deep learning model, batch prediction of soil and water loss for all pixels in the region is completed in parallel, ultimately achieving rapid aggregation of the total regional amount and visualization of the spatial distribution of risk.

[0079] The method for predicting soil erosion based on rainfall events provided in this embodiment can automatically drive parallel calculations of dedicated models for different natural plots by synchronously inputting forecast rainfall data and high-resolution spatial attribute data. While ensuring that the prediction results strictly match the actual conditions of each plot of land, it can complete batch predictions for the entire region at once, significantly improving the timeliness, accuracy, and operational capability of the predictions.

[0080] As one or more specific application embodiments of the present invention, combined with Figures 5 to 6 The method for predicting soil erosion based on rainfall events provided by this invention will be further described in detail. Taking a small watershed as an example, the specific process for predicting and analyzing soil erosion based on rainfall events is as follows: (1) Training data and model training: To prepare historical rainfall and soil erosion monitoring data for a small watershed, the data needs to include "station name," "rainfall," "rainfall intensity," "runoff per unit area," and "scour per unit area." The historical monitoring data will be categorized by slope (5°–15°, 15°–20°, 20°–25°, and above 25°), vegetation cover (below 30%, 30%–60%, 60%–80%, and above 80%), and the type of soil and water conservation engineering measures. A deep learning model will be used to complete the data by filling in missing values ​​and deleting combinations with fewer than 20 data points. The data will then be divided into a 7:1.5:1.5 training set, validation set, and test set. The model will be trained and self-validated based on the categorized data to obtain the trained model combination.

[0081] The basic information of the runoff plots is shown in Table 1 below, and the data classification for slopes above 25° and vegetation coverage above 80% is shown in Table 2 below.

[0082] Table 1 Basic Information on Runoff Zones

[0083] Table 2 Data Classification

[0084] (2) Geospatial data preparation and natural land parcel matching: Based on land use data, artificially hardened areas such as villages and roads are removed from the predicted area. Then, based on the small watershed digital elevation model and normalized difference vegetation index data, the slope and vegetation cover of the small watershed are calculated respectively. Combined with the soil and water conservation engineering measures data, the site characteristics of natural plots and runoff plots are matched and divided to obtain the divided plot data.

[0085] (3) Rainfall data processing: The required rainfall forecast data preparation includes two formats: The first is a uniform rainfall model, which reads hourly rainfall sequences in TXT text format; this format is generally suitable for situations where rainfall characteristics are consistent across the forecast area. The second is a spatially distributed rainfall model, which reads raster sequences of rainfall distribution maps in GeoTIFF format; this format is suitable for situations where rainfall is unevenly distributed in time and space within the forecast area. The first type of rainfall data can be directly input into the model for prediction, while the second type can be correlated with the spatial data of divided land parcels to support accurate predictions within different parcels.

[0086] (4) Simulation prediction and result display: The forecasted rainfall data is input into the trained deep learning model. The model automatically obtains rainfall amount and rainfall intensity information. At the same time, based on the input land use, slope, vegetation cover and soil and water conservation engineering measures data, it completes the matching and division of natural plots and carries out the prediction of soil and water loss during each rainfall event. The prediction results include the slope runoff and soil erosion of the predicted area under the corresponding rainfall event conditions, and generate a spatial distribution raster map.

[0087] Example of prediction results: The prediction of soil erosion during the entire rainfall event is complete. The rainfall lasted for 13 hours, with a rainfall of 100 mm and a maximum hourly rainfall intensity of 18.2 mm. The predicted results are a slope runoff of 70,320.3 cubic meters and a soil erosion of 140.7 tons.

[0088] The predicted results for slope runoff and soil erosion are as follows: Figure 5 and Figure 6 As shown.

[0089] The method for predicting soil erosion based on rainfall events in this invention sets classification rules for runoff plot site conditions according to different slope grades, vegetation cover, and engineering measure types. It collects and organizes historical typical rainfall event monitoring data and corresponding soil erosion monitoring data for each category of site conditions, constructing a multi-factor combined training dataset for rainfall event slope runoff and soil erosion. A deep learning method is used to construct a soil erosion prediction model, capturing the temporal characteristics of rainfall-runoff-erosion through sequence learning, and establishing a nonlinear mapping relationship from rainfall event characteristics to slope runoff and soil erosion under different site conditions. A natural plot matching and partitioning method based on site characteristics is established. Hardened areas such as villages and roads in the land use data are removed, and then the slope, vegetation cover, and soil and water conservation engineering measure data within the prediction area are spatially matched to complete the natural plot matching and partitioning based on different site conditions. Using forecasted rainfall as input, and combining digital elevation models, vegetation indices, and engineering data, a trained deep learning model is used to perform batch predictions of slope runoff and soil erosion for various natural plots under different rainfall events. This allows for the prediction of slope runoff and soil erosion across a region. This forecasting method enables soil and water loss prediction and analysis driven by forecasted rainfall events, providing technical support for soil and water conservation management decisions.

[0090] This embodiment also provides a rainfall-based soil erosion prediction device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0091] This embodiment provides a device for predicting soil erosion during a single rainfall event, such as... Figure 7 As shown, it includes: The multi-factor combined dataset construction module 701 is used to set classification rules for runoff plot site conditions according to different slope levels, vegetation cover and engineering measure types, obtain historical rainfall data and soil erosion monitoring data of runoff plots with different site condition types in the target area or similar topographic and geomorphological areas, and construct a multi-factor combined dataset based on the historical rainfall data and soil erosion monitoring data.

[0092] The model training module 702 is used to train a preset deep learning model based on a multi-factor combination dataset to obtain a rainfall-based soil erosion prediction model.

[0093] The soil and water loss prediction module 703 is used to divide the target area into natural plots with different site conditions according to classification rules, and drive the precipitation soil and water loss prediction model to predict precipitation soil and water loss for each natural plot based on the collected forecast rainfall data.

[0094] In some optional implementations, the multi-factor combination dataset construction module 701 includes: The data classification unit is used to classify the site conditions of runoff plots based on different slope grades, vegetation cover and engineering measures. It classifies historical rainfall data and soil erosion monitoring data according to site condition types.

[0095] In some optional implementations, historical rainfall data includes rainfall characteristics, and soil erosion monitoring data includes slope runoff and soil erosion; the multi-factor combined dataset construction module 701 also includes: The multi-factor combined dataset construction unit is used to construct input-output data pairs based on the site condition types of the input and output data, using the characteristics of each rainfall event as input data and the slope runoff and soil erosion as output data. The input-output data pairs bound to the site condition types are used as multi-factor combined datasets.

[0096] In some alternative implementations, the model training module 702 includes: The multi-factor combination dataset is input into the preset deep learning model, which captures the features of rainfall-runoff-erosion through sequence learning and establishes a nonlinear mapping relationship between the features of each rainfall event and the slope runoff and soil erosion under different site conditions, thus obtaining a rainfall-soil loss prediction model.

[0097] In some alternative implementations, the soil erosion prediction module 703 includes: The natural surface extent determination unit is used to acquire land use data of the target area and, based on the land use data, remove hardened surface areas such as villages and roads to obtain the natural surface extent to be delineated.

[0098] The multi-source spatial data generation unit is used to acquire digital elevation data, vegetation data, and soil and water conservation engineering measures of the target area, and based on the digital elevation data, vegetation data, and soil and water conservation engineering measures, generate multi-source spatial data with different slope levels, vegetation cover levels, and soil and water conservation engineering measure types corresponding to preset classification rules.

[0099] Natural plot generation unit is used to unify multi-source spatial data into the same geographic coordinate system and spatial resolution to obtain natural plots with different site conditions.

[0100] In some optional implementations, the soil erosion prediction module 703 further includes: The prediction unit is used to take forecast rainfall data, different slope grades, vegetation cover grades and types of soil and water conservation engineering measures as inputs, drive the soil and water loss prediction model to make batch predictions of slope runoff and soil erosion of each natural plot under the conditions of each rainfall event, and obtain the soil and water loss prediction results of each natural plot in the target area.

[0101] In some optional implementations, the forecast rainfall data supports the following input methods: Read uniform rainfall patterns from hourly rainfall sequences in text format or spatially distributed rainfall patterns from rainfall raster sequences in GeoTIFF format.

[0102] The rainfall-based soil erosion prediction device provided in this embodiment of the invention can execute the rainfall-based soil erosion prediction method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0103] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0104] The following is a detailed reference. Figure 8 This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 801, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 802 or a program loaded from memory 808 into random access memory (RAM) 803. The RAM 803 also stores various programs and data required for the operation of the electronic device. The processor 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0105] Typically, the following devices can be connected to I / O interface 805: input devices 806 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 807 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 808 including, for example, magnetic tapes, hard disks, etc.; and communication devices 809. Communication device 809 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 8 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0106] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 809, or installed from a memory 808, or installed from a ROM 802. When the computer program is executed by the processor 801, it performs the functions defined in the event-based rainfall-based soil erosion prediction method of the embodiments of the present invention.

[0107] Figure 8 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0108] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the method for predicting soil erosion during rainfall as shown in the above embodiments is implemented.

[0109] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0110] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and all such modifications and variations fall within the scope defined by the appended claims.

Claims

1. A method for predicting event-based rainfall soil erosion, characterized by, The method comprises: setting classification rules for the site conditions of the runoff plots according to different slope grades, vegetation cover degrees and engineering measure types; obtaining historical rainfall data and soil erosion monitoring data of runoff plots of different site condition types in the target region or a similar topography region, and constructing a multi-factor combined data set based on the historical rainfall data and soil erosion monitoring data; training a preset deep learning model based on the multi-factor combined data set to obtain a rainfall event soil erosion prediction model; based on the classification rules, dividing the target region into natural blocks of different site conditions, and based on the collected forecast rainfall data, driving the rainfall event soil erosion prediction model to predict the rainfall event soil erosion of each natural block.

2. The method of claim 1, wherein, The method comprises: based on the classification rules for the site conditions of the runoff plots classified according to different slope grades, vegetation cover degrees and engineering measure types, classifying the historical rainfall data and soil erosion monitoring data according to site condition types.

3. The method of claim 1, wherein, The historical rainfall data includes rainfall event characteristics, and the soil erosion monitoring data includes slope runoff and soil erosion amount; based on the historical rainfall data and soil erosion monitoring data, constructing a multi-factor combined data set comprises: taking the rainfall event characteristics as input data, taking the slope runoff and soil erosion amount as output data, and based on the site condition types to which the input data and output data belong, constructing input-output data pairs, and taking the input-output data pairs bound with site condition types as the multi-factor combined data set.

4. The rainfall event soil erosion prediction method according to claim 1, wherein based on the multi-factor combined data set, training a preset deep learning model to obtain a rainfall event soil erosion prediction model comprises: inputting the multi-factor combined data set into the preset deep learning model, so that the preset deep learning model captures the time sequence characteristics of rainfall-runoff-erosion through sequence learning, and establishes a nonlinear mapping relationship from the rainfall event characteristics to the slope runoff and soil erosion amount of different types of site conditions, to obtain the rainfall event soil erosion prediction model.

5. The method of claim 1, wherein, based on the preset classification rules, dividing the target region into natural blocks of different site conditions comprises: obtaining land use data of the target region, and based on the land use data, excluding the hardened ground surface area of villages and roads to obtain the natural ground surface range to be divided; obtaining digital elevation data, vegetation data and soil and water conservation engineering measure data of the target region, and based on the digital elevation data, vegetation data and soil and water conservation engineering measure data, generating multi-source spatial data of different slope grades, vegetation cover degrees and soil and water conservation engineering measure types corresponding to the classification rules; unifying the multi-source spatial data into the same geographic coordinate system and spatial resolution to obtain natural blocks of different site conditions.

6. The method of claim 1, wherein, The predicted rainfall data is collected, and the soil erosion prediction model is driven to predict the field rainfall soil erosion of each natural plot. The predicted rainfall data, different slope levels, vegetation coverage levels, and water and soil conservation engineering measure types are taken as inputs to drive the soil erosion prediction model to batch predict the slope runoff and soil erosion amount of each natural plot under field rainfall conditions, and the soil erosion prediction results of each natural plot in the target region are obtained.

7. The field rainfall soil loss prediction method according to claim 1 or 6, characterized by, The predicted rainfall data supports the following input modes: Reading a uniform rainfall pattern of a text format hourly rainfall sequence or reading a spatial distribution rainfall pattern of a GeoTIFF format rainfall raster sequence.

8. A device for predicting soil erosion by rainfall for a game, characterized by, The device comprises: A multi-factor combination dataset construction module configured to set classification rules for the runoff plot site conditions according to different slope levels, vegetation coverage, and engineering measure types, obtain historical field rainfall data and soil erosion monitoring data of runoff plots of different site condition types in the target region or a similar topography region, and construct a multi-factor combination dataset based on the historical field rainfall data and soil erosion monitoring data; A model training module configured to train a preset deep learning model based on the multi-factor combination dataset to obtain a field rainfall soil erosion prediction model; A soil erosion prediction module configured to divide the target region into natural plots of different site conditions according to a preset classification rule, and drive the field rainfall soil erosion prediction model to predict the field rainfall soil erosion of each natural plot based on the collected predicted rainfall data.

9. An electronic device, comprising: It comprises: A memory and a processor, which are communicatively connected, and the memory stores computer instructions, and the processor executes the computer instructions to perform the field rainfall soil erosion prediction method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for causing a computer to execute the field rainfall soil erosion prediction method of any one of claims 1 to 7.

Images