Drought forest fire disaster-causing element extraction and identification method and system, storage medium and computer program product

Abstract

The invention discloses a drought forest fire disaster-inducing element extraction and identification method and system, a storage medium and a computer program product, and relates to the technical field of remote sensing disaster monitoring, and the method comprises the steps: collecting a multi-source data set of a drought forest fire region, and extracting disaster-inducing factors from the multi-source data, the multi-source data set comprising satellite data; performing time-space two-dimensional thermal anomaly recognition based on the satellite data, and obtaining a fire point recognition result in combination with a drought forest fire disaster chain anomaly index output by a pre-constructed drought forest fire disaster chain anomaly index extraction model; the drought forest fire disaster chain anomaly index extraction model is constructed based on a drought forest fire disaster chain knowledge model, and the drought forest fire disaster chain knowledge model is obtained based on spatial and temporal distribution relation analysis of disaster-inducing factors. Through collection of multi-source data, construction of a knowledge model and an anomaly index extraction model and space-time two-dimensional thermal anomaly identification, efficient identification and accurate extraction of disaster-causing elements of a drought forest fire disaster chain are realized.

Description

Technical Field

[0001] This application relates to the field of remote sensing disaster monitoring technology, and in particular to methods, systems, storage media and computer program products for extracting and identifying drought and forest fire disaster-causing factors. Background Technology

[0002] The drought-forest fire disaster chain is a complex natural disaster process driven by the coupling of multiple factors such as meteorological conditions, underlying surface characteristics, and vegetation status. With the development of remote sensing technology, multi-source remote sensing data from meteorological satellites, high-resolution optical satellites, and radar satellites provide new ideas for drought-forest fire disaster monitoring. However, due to the different data formats of various sensors, the accuracy varies greatly, and there is a lack of quantitative modeling of the disaster chain evolution mechanism. As a result, the application of multi-source remote sensing data in drought-forest fire monitoring still faces problems such as difficulty in data fusion, weak correlation of disaster-causing factors, and low efficiency in extracting abnormal features. This leads to insufficient accuracy in identifying key disaster-causing factors, slow response speed, and an inability to effectively support early warning and emergency decision-making for drought-forest fire disaster chains.

[0003] Therefore, how to utilize multi-source data to achieve efficient identification and accurate extraction of disaster-causing factors in the drought and forest fire disaster chain has become a technical problem that this application urgently needs to solve.

[0004] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a method, system, storage medium, and computer program product for extracting and identifying disaster-causing factors of drought and forest fire, aiming to solve the technical problem of how to use multi-source data to achieve efficient identification and accurate extraction of disaster-causing factors in the drought and forest fire disaster chain.

[0006] To achieve the above objectives, this application proposes a method for extracting and identifying drought-induced forest fire disaster factors, the method comprising: Collect multi-source datasets of drought and forest fire areas, and extract disaster-causing factors from the multi-source datasets, which include satellite data; Based on the satellite data, spatiotemporal dual-dimensional thermal anomaly identification is performed, and combined with the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model, the fire point identification result is obtained; wherein, the drought and forest fire disaster chain anomaly index extraction model is constructed based on the drought and forest fire disaster chain knowledge model, which is obtained based on the spatiotemporal distribution relationship analysis of the disaster-causing factors.

[0007] In one embodiment, before the step of performing spatiotemporal dual-dimensional thermal anomaly identification based on the satellite data and combining it with the drought and forest fire disaster chain anomaly index output by a pre-constructed drought and forest fire disaster chain anomaly index extraction model to obtain the fire point identification result, the method further includes: Analyze the spatiotemporal distribution relationship between the disaster-causing factors and the drought-stricken forest fire areas, and construct a drought-stricken forest fire disaster chain knowledge model based on the spatiotemporal distribution relationship; Based on the drought and forest fire disaster chain knowledge model and the disaster-causing factors, an abnormal index extraction model for the drought and forest fire disaster chain is constructed.

[0008] In one embodiment, the step of analyzing the spatiotemporal distribution relationship between the disaster-causing factors and the drought-stricken forest fire area, and constructing a drought-stricken forest fire disaster chain knowledge model based on the spatiotemporal distribution relationship includes: Statistical methods were used to extract the temporal dynamics and spatial differentiation characteristics of the disaster-causing factors and the drought-stricken forest fire areas, respectively. A spatiotemporal distribution relationship is constructed based on the aforementioned temporal dynamic characteristics and spatial differentiation characteristics; A knowledge model of drought and forest fire disaster chain is constructed based on the spatiotemporal distribution relationship.

[0009] In one embodiment, the step of constructing a drought and forest fire disaster chain anomaly index extraction model based on the drought and forest fire disaster chain knowledge model and the disaster-causing factors includes: The hold-out method is used to divide the disaster-causing factors into a training set and a test set; Based on the training set and the drought and forest fire disaster chain knowledge model, an adaptation algorithm is selected and trained to obtain an initial drought and forest fire disaster chain anomaly index extraction model. The initial drought-forest fire disaster chain anomaly index extraction model was validated and optimized using the test set, resulting in the drought-forest fire disaster chain anomaly index extraction model.

[0010] In one embodiment, the step of collecting multi-source datasets of drought-stricken forest fire areas and extracting disaster-causing factors from the multi-source data includes: Multi-source datasets from arid forest fire areas were collected, and the multi-source datasets were standardized to obtain standardized datasets. Factors related to drought processes are extracted from meteorological data in the standardized dataset, and factors related to drought and forest fire disasters are extracted from satellite data in the standardized dataset. Disaster-causing factors are generated based on the drought process-related factors and the drought-forest fire disaster-related factors, combined with the historical wildfire event catalog in the standardized dataset.

[0011] In one embodiment, the steps of performing spatiotemporal dual-dimensional thermal anomaly identification based on the satellite data and combining it with the drought and forest fire disaster chain anomaly index output by a pre-constructed drought and forest fire disaster chain anomaly index extraction model to obtain the fire point identification result include: The satellite data is preprocessed to obtain a standardized satellite brightness temperature dataset; The drought and forest fire disaster chain anomaly index extraction model is invoked, and the real-time disaster-causing factors are input into the drought and forest fire disaster chain anomaly index extraction model to obtain the drought and forest fire disaster chain anomaly index. Based on the drought and forest fire disaster chain anomaly index, the satellite brightness temperature dataset is used to identify thermal anomalies in a spatiotemporal dimension to obtain fire point identification results. The spatiotemporal dual-dimensional thermal anomaly identification includes absolute fire point discrimination, conditional fire point discrimination, low temperature fire point monitoring, continuous fire judgment, flare spot removal, and underlying surface judgment.

[0012] In one embodiment, the step of combining the drought and forest fire disaster chain anomaly index to perform spatiotemporal dual-dimensional thermal anomaly point identification on the satellite brightness temperature dataset to obtain fire point identification results includes: By combining the drought forest fire disaster chain anomaly index with the satellite brightness temperature dataset, absolute fire point identification is performed to obtain preliminary absolute fire point identification results; Conditional fire point discrimination is performed on the satellite brightness temperature dataset from both temporal and spectral dimensions to obtain preliminary conditional fire point identification results; the preliminary conditional fire point identification results include conditional fire point pixels and brightness temperature features corresponding to the conditional fire point pixels. The system captures minute temperature rise signals of the conditional fire point pixels to perform low-temperature fire point monitoring, and monitors the conditional fire point pixels from a time dimension to determine continuous fires. The preliminary identification results of absolute fire points are merged and deduplicated with the conditional fire point pixels obtained by the low-temperature fire point monitoring and the continuous fire judgment to obtain a fire point pixel set. Flare spots are removed from the fire point pixel set to obtain the fire point candidate set; The underlying surface type is verified on the candidate fire point set to obtain the fire point identification result.

[0013] Furthermore, to achieve the above objectives, this application also proposes a drought- and forest fire-causing factor extraction and identification system, which includes: The data collection module is used to collect multi-source datasets of drought and forest fire areas and extract disaster-causing factors from the multi-source datasets, which include satellite data. The thermal anomaly identification module is used to identify thermal anomalies in a spatiotemporal dual dimension based on the satellite data, and to obtain the fire point identification result by combining the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model; wherein, the drought and forest fire disaster chain anomaly index extraction model is constructed based on the drought and forest fire disaster chain knowledge model, which is obtained by analyzing the spatiotemporal distribution relationship of the disaster-causing factors.

[0014] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the drought and forest fire disaster-causing element extraction and identification method described above.

[0015] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the drought and forest fire disaster-causing element extraction and identification method described above.

[0016] One or more technical solutions proposed in this application have at least the following technical effects: A multi-source dataset of drought-stricken forest fire areas is collected, and disaster-causing factors are extracted from the multi-source dataset, which includes satellite data. Spatiotemporal dual-dimensional thermal anomaly identification is performed based on the satellite data, and the drought-stricken forest fire disaster chain anomaly index output by a pre-constructed drought-stricken forest fire disaster chain anomaly index extraction model is combined to obtain the fire point identification result. The drought-stricken forest fire disaster chain anomaly index extraction model is constructed based on a drought-stricken forest fire disaster chain knowledge model, which is obtained by analyzing the spatiotemporal distribution relationship of the disaster-causing factors. First, the collection of multi-source datasets provides a comprehensive data foundation for the efficient identification and accurate extraction of disaster-causing factors in the drought-forest fire disaster chain, ensuring multi-dimensional coverage of disaster-causing factors. Furthermore, spatiotemporal distribution relationship analysis of disaster-causing factors is conducted to construct a knowledge model, clarifying the core objectives for subsequent anomaly identification and improving extraction accuracy. Further, the construction of an anomaly index extraction model for the drought-forest fire disaster chain enables efficient calculation and accurate modeling of the abnormal characteristics of disaster-causing factors. Finally, spatiotemporal dual-dimensional thermal anomaly identification achieves rapid location of fire points and interference elimination from both temporal and spatial dimensions, ultimately achieving efficient identification and accurate extraction of disaster-causing factors in the drought-forest fire disaster chain. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the first embodiment of the drought and forest fire hazard extraction and identification method of this application; Figure 2 A flowchart illustrating the second embodiment of the drought and forest fire hazard extraction and identification method of this application; Figure 3 A flowchart illustrating the third embodiment of the drought and forest fire hazard extraction and identification method of this application; Figure 4 A simplified flowchart illustrating the drought and forest fire hazard extraction and identification method provided in this application embodiment; Figure 5 This is a schematic diagram of the module structure of the drought and forest fire disaster-causing element extraction and identification device according to an embodiment of this application; Figure 6 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the drought and forest fire disaster-causing element extraction and identification method in the embodiments of this application.

[0020] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0022] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution of this application embodiment is: to collect multi-source datasets of drought and forest fire areas and extract disaster-causing factors from the multi-source datasets, the multi-source datasets including satellite data; to perform spatiotemporal dual-dimensional thermal anomaly identification based on the satellite data, and to obtain fire point identification results by combining the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model; wherein, the drought and forest fire disaster chain anomaly index extraction model is constructed based on the drought and forest fire disaster chain knowledge model, and the drought and forest fire disaster chain knowledge model is obtained based on the spatiotemporal distribution relationship analysis of the disaster-causing factors.

[0024] This application takes into account that the drought-forest fire disaster chain is a complex natural disaster process driven by multiple factors such as meteorological conditions, underlying surface characteristics, and vegetation status. With the development of remote sensing technology, multi-source remote sensing data from meteorological satellites, high-resolution optical satellites, and radar satellites provide new ideas for drought-forest fire disaster monitoring. However, due to the different data formats of different sensors, the accuracy varies greatly, and there is a lack of quantitative modeling of the disaster chain evolution mechanism. As a result, the application of multi-source remote sensing data in drought-forest fire monitoring still faces problems such as difficulty in data fusion, weak correlation of disaster-causing factors, and low efficiency in extracting abnormal features. This leads to insufficient accuracy in identifying key disaster-causing factors, slow response speed, and an inability to effectively support early warning and emergency decision-making for drought-forest fire disaster chains.

[0025] Therefore, this application provides a solution to collect multi-source datasets of drought-stricken forest fire areas and extract disaster-causing factors from the multi-source datasets, including satellite data; perform spatiotemporal dual-dimensional thermal anomaly identification based on the satellite data, and combine the drought-stricken forest fire disaster chain anomaly index output by a pre-constructed drought-stricken forest fire disaster chain anomaly index extraction model to obtain fire point identification results; wherein, the drought-stricken forest fire disaster chain anomaly index extraction model is constructed based on a drought-stricken forest fire disaster chain knowledge model, and the drought-stricken forest fire disaster chain knowledge model is obtained based on the spatiotemporal distribution relationship analysis of the disaster-causing factors. First, the collection of multi-source datasets provides a comprehensive data foundation for the efficient identification and accurate extraction of disaster-causing factors in the drought-forest fire disaster chain, ensuring multi-dimensional coverage of disaster-causing factors. Furthermore, spatiotemporal distribution relationship analysis of disaster-causing factors is conducted to construct a knowledge model, clarifying the core objectives for subsequent anomaly identification and improving extraction accuracy. Further, the construction of an anomaly index extraction model for the drought-forest fire disaster chain enables efficient calculation and accurate modeling of the abnormal characteristics of disaster-causing factors. Finally, spatiotemporal dual-dimensional thermal anomaly identification achieves rapid location of fire points and interference elimination from both temporal and spatial dimensions, ultimately achieving efficient identification and accurate extraction of disaster-causing factors in the drought-forest fire disaster chain.

[0026] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, such as a drought and forest fire causative factor extraction and identification system. The following description uses a drought and forest fire causative factor extraction and identification system as an example to illustrate this embodiment and the subsequent embodiments.

[0027] Based on this, embodiments of this application provide a method for extracting and identifying drought- and forest fire-causing factors, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the drought and forest fire disaster-causing factor extraction and identification method of this application.

[0028] In this embodiment, the method for extracting and identifying drought and forest fire disaster-causing factors includes steps S10 to S20: Step S10: Collect multi-source datasets of drought and forest fire areas, and extract disaster-causing factors from the multi-source datasets, which include satellite data; It should be noted that, in the embodiments of this application, the drought-prone forest fire area refers to a forest or grassland area where the vegetation moisture content has decreased due to prolonged drought, and the risk of fire has increased significantly.

[0029] Multi-source datasets refer to data collections collected through multiple channels for analyzing the causes of disasters. In addition to satellite data, they also include historical wildfire event catalogs (databases that record the time, location, and scale of past fires) and meteorological data (such as observational data reflecting climate conditions, such as temperature, precipitation, and relative humidity).

[0030] Disaster-causing factors refer to key elements that may trigger or exacerbate drought and forest fire disasters, including meteorological factors (such as Standardized Precipitation Index (SPI), daily average evaporation, and drought duration), underlying surface factors (such as vegetation drought index, combustible moisture content, and vegetation cover type) and surface temperature.

[0031] The core objective of this step is to systematically extract key factors influencing drought and forest fire disasters by integrating multi-source data, providing fundamental data support for subsequent disaster chain analysis and fire point identification. The approach is as follows: First, data is collected through various means, including satellite remote sensing and ground observation stations, ensuring data coverage across multiple dimensions such as meteorology, vegetation, and topography. Then, the raw data is preprocessed (e.g., radiometric correction, cloud removal, format conversion) to eliminate noise and errors. Finally, disaster-causing factors are extracted based on data characteristics; for example, the vegetation drought index is retrieved from satellite data, and standardized precipitation factors and drought duration days are calculated from meteorological data. This step enables a multi-dimensional characterization of disaster causes, avoiding the limitations of single data sources and improving the comprehensiveness and accuracy of disaster-causing factor extraction.

[0032] In one possible implementation, satellite data can be selected from remote sensing data of different resolutions or types, such as high-resolution optical satellites for extracting land cover types with an accuracy of 30 meters, and geostationary meteorological satellites for acquiring real-time brightness and temperature data; meteorological data can be integrated with ground meteorological station observation data and reanalysis data to compensate for the problem of uneven distribution of observation stations.

[0033] In another possible implementation, the extraction of hazard factors can be carried out using automated algorithms, such as directly inverting the moisture content of combustibles from the raw spectral data based on machine learning models, replacing traditional empirical formula calculations and improving extraction efficiency.

[0034] Step S20: Based on the satellite data, perform spatiotemporal dual-dimensional thermal anomaly identification, and combine the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model to obtain the fire point identification result; wherein, the drought and forest fire disaster chain anomaly index extraction model is constructed based on the drought and forest fire disaster chain knowledge model, and the drought and forest fire disaster chain knowledge model is obtained based on the spatiotemporal distribution relationship analysis of the disaster-causing factors.

[0035] It should be noted that, in the embodiments of this application, spatiotemporal dual-dimensional thermal anomaly identification refers to a method for detecting surface high temperature anomaly areas by combining time series changes and spatial pixel differences. The "time dimension" focuses on the dynamic changes of pixel temperature over time, while the "spatial dimension" analyzes the temperature differences between the target pixel and surrounding pixels.

[0036] The drought and forest fire disaster chain anomaly index extraction model refers to a model constructed by an algorithm based on the correlation between disaster-causing factors and fires, which is used to quantify the degree of disaster risk anomalies. Its output, the "drought and forest fire disaster chain anomaly index", is a quantitative indicator that comprehensively reflects the anomaly level of disaster-causing factors.

[0037] The drought-forest fire disaster chain knowledge model refers to a theoretical model that describes the mechanism of disaster chain occurrence and development by analyzing the spatiotemporal distribution patterns of disaster-causing factors and forest fires (such as the positive correlation between the number of drought days and the frequency of fire occurrence).

[0038] The core objective of this step is to achieve accurate identification of fire points by combining thermal anomaly recognition with an anomaly index model, especially distinguishing real fire points from interfering factors (such as solar flares or high temperatures in non-forested areas). The approach is as follows: First, perform spatiotemporal dual-dimensional thermal anomaly detection based on satellite data. Suspected fire points are initially screened using absolute fire point discrimination (e.g., fixed brightness temperature threshold) and conditional fire point discrimination (e.g., temperature change rate and band differences). Then, the screening results are combined with the risk index output by the anomaly index model. For example, if the anomaly index is higher than a threshold and the thermal anomaly meets the fire point conditions, it is determined to be a real fire point. This step effectively reduces the false positive rate, improves the accuracy and reliability of fire point identification, and enables early monitoring of low-temperature fire points and small fires.

[0039] For example, in one specific implementation, based on the disaster-causing factors extracted from multi-source datasets in arid forest fire areas, traditional statistical analysis reveals that when the vegetation drought index > 0.6 and SPI-3 < -1.5, the probability of forest fire occurrence significantly increases. Based on this, a knowledge model of the drought forest fire disaster chain is constructed. Subsequently, a Bayesian probabilistic algorithm is used to construct an anomaly index extraction model. 80% of the historical disaster-causing factor data from the multi-source dataset's historical wildfire event catalog is used as the training set, and 20% as the test set. The output is an anomaly index ranging from 0 to 1 (a higher value indicates a more abnormal disaster risk). In thermal anomaly identification, FY-4A satellite data is used for judgment: when the brightness temperature of the B7 band > 360K, it is directly determined as an absolute fire point; when the pixel temperature rise within 10 minutes > 5K and the brightness temperature difference between the B7 and B12 bands > 10K (20K during the day), it is determined as a conditional fire point. Finally, the thermal anomaly results are combined with the anomaly index: if the anomaly index is >0.8 and the thermal anomaly is determined to be a fire point, and the underlying surface is forest-grassland (based on 30-meter land cover data), then the fire point identification result is output.

[0040] It should be noted that SPI-3 represents the 3-month standardized precipitation index, a drought monitoring index obtained by normalizing the cumulative precipitation over the past 3 months through probability distribution. Its value typically ranges from -3.0 (extreme drought) to 3.0 (extreme wetness). SPI-3 is mainly used to assess seasonal drought conditions and is one of the key disaster-causing factors in agricultural drought and forest fire risk early warning, reflecting the long-term trend of surface combustible moisture content.

[0041] FY-4A is the second-generation geostationary meteorological satellite "Fengyun-4A". As a geostationary orbit satellite, FY-4A carries payloads such as the multi-channel scanning imaging radiometer (AGRI), which can provide high temporal resolution (minute-level) and multispectral band (including visible light, infrared, and thermal infrared) observation data. It is widely used in weather monitoring, disaster early warning (such as forest fires and droughts) and climate research. Its thermal infrared band data is the core data source for brightness temperature inversion in fire point identification.

[0042] B7 band brightness temperature refers to the brightness temperature of the 7th band (Band 7) of a satellite sensor. "Band 7" is a specific spectral channel of a satellite sensor (e.g., the B7 band of the FY-4A AGRI sensor is typically in the mid-infrared or thermal infrared band, with a wavelength range of approximately 3.5-4.0 μm). "Brightness temperature" is the equivalent blackbody temperature (unit: Kelvin, K) obtained by inverting the radiance values ​​of this band. It reflects the thermal radiation intensity of a target object (such as the Earth's surface or a fire point) in this band and is a fundamental parameter for identifying high-temperature anomalies in fire point detection.

[0043] The brightness-temperature difference in the B12 band refers to the difference in brightness temperature between Band 12 of a satellite sensor and other bands (usually adjacent thermal infrared bands). The B12 band is generally a long-wave thermal infrared band (e.g., the wavelength of the B12 band in the FY-4A AGRI sensor is approximately 10.3-11.3 μm). The "brightness-temperature difference" refers to the difference in brightness temperature between the B12 band and another band (e.g., B7 or B11 bands) (e.g., B12-B7). The radiation characteristics of fire points differ significantly across different thermal infrared bands (e.g., high-temperature fire points have higher brightness temperatures in the mid-infrared band). The brightness-temperature difference can be used to distinguish real fire points from background thermal noise (e.g., flares, high-temperature underlying surfaces).

[0044] 360K, 10K, 20K, and 5K are all specific values ​​of the temperature unit "Kelvin (K)," which are usually used as the brightness temperature threshold or brightness temperature difference threshold in fire point identification.

[0045] This embodiment provides a method for extracting and identifying disaster-causing factors in drought and forest fire disaster chains. The collection of multi-source datasets provides a comprehensive data foundation for the efficient identification and accurate extraction of disaster-causing factors in the drought and forest fire disaster chain, ensuring multi-dimensional coverage of disaster-causing factors. Furthermore, spatiotemporal distribution relationship analysis of disaster-causing factors is conducted to construct a knowledge model, clarifying the core objectives for subsequent anomaly identification and improving extraction accuracy. Further, the construction of an anomaly index extraction model for the drought and forest fire disaster chain enables efficient calculation and accurate modeling of the abnormal characteristics of disaster-causing factors. Finally, spatiotemporal dual-dimensional thermal anomaly identification achieves rapid location of fire points and interference elimination from both temporal and spatial dimensions, ultimately achieving efficient identification and accurate extraction of disaster-causing factors in the drought and forest fire disaster chain.

[0046] Based on the first embodiment of this application, a second embodiment of this application is proposed. In the second embodiment of this application, content that is the same as or similar to that in the first embodiment described above can be referred to the above description and will not be repeated hereafter.

[0047] Based on this, please refer to Figure 2 , Figure 2 The flowchart provided for the second embodiment of this application is as follows: Figure 2 As shown, before step S20, the method for extracting and identifying drought and forest fire disaster-causing factors further includes steps S01 to S02: Step S01: Analyze the spatiotemporal distribution relationship between the disaster-causing factors and the drought-stricken forest fire area, and construct a drought-stricken forest fire disaster chain knowledge model based on the spatiotemporal distribution relationship; It should be noted that, in the embodiments of this application, the spatiotemporal distribution relationship refers to the correlation between the temporal dynamic changes of disaster-causing factors (such as seasonal fluctuations and interannual differences) and spatial differentiation characteristics (such as distribution differences under different terrains and vegetation types) and the occurrence of forest fires.

[0048] The core objective of this step is to construct a quantitative theoretical model by systematically analyzing the spatiotemporal correlation between disaster-causing factors and forest fires, providing a scientific basis for subsequent anomaly index extraction. The approach is as follows: First, analyze the dynamic changes of disaster-causing factors from a temporal perspective (e.g., the monthly correlation between drought duration and fire frequency); second, analyze the differentiation characteristics of factors from a spatial perspective (e.g., the relationship between vegetation drought index and fire density at different altitudes); third, integrate spatiotemporal characteristics to reveal the coupling mechanism of "factor anomaly-fire response"; finally, refine the patterns into a computable knowledge model, such as "when SPI < -1.5 and vegetation drought index > 0.6, the probability of forest fire occurrence increases by 3 times." This step transforms the complex disaster chain process into quantifiable rules, laying a theoretical foundation for model construction.

[0049] In one possible implementation, the spatiotemporal distribution relationship analysis can introduce a spatiotemporal weighted regression model, which can improve the ability to capture local patterns by assigning different weights to samples at different spatiotemporal locations; the knowledge model construction can adopt an ontological approach, which formally defines disaster-causing factors, spatiotemporal characteristics, fire response and other elements as ontological concepts and relationships, thereby enhancing the interpretability of the model.

[0050] In another possible implementation, geographic information system visualization technology can be used to visually display the spatial superposition characteristics of disaster-causing factors and forest fires, and to assist in the extraction of patterns.

[0051] Step S02: Construct an abnormal index extraction model for the drought and forest fire disaster chain based on the knowledge model of the drought and forest fire disaster chain and the disaster-causing factors.

[0052] In order to transform the qualitative rules of the knowledge model into a computable quantitative model and realize the automated identification and quantitative assessment of the abnormal state of disaster-causing factors, this implementation method proposes, specifically: First, the mapping relationship between the model input and output is determined based on the rules in the knowledge model. The model input consists of real-time disaster-causing factors in forest fire and drought areas collected by remote sensing satellites, and the model output is the anomaly index. According to the specific mapping relationship, an appropriate algorithm (such as Bayesian probability or random forest) is selected to embed the knowledge rules into the model training process. Finally, through training and validation on the dataset, the model can output the anomaly index based on the real-time disaster-causing factor data. For example, when the input SPI = -1.8 and the vegetation drought index = 0.7, the model outputs an anomaly index of 0.85 (high anomaly).

[0053] By constructing an abnormal index extraction model for drought and forest fire disaster chains, abstract knowledge rules can be transformed into predictive tools, providing quantitative risk references for fire point identification.

[0054] In this embodiment, by revealing the spatiotemporal distribution patterns of disaster-causing factors and forest fires, the constructed knowledge model clarifies the threshold conditions and spatiotemporal coupling mechanisms of key disaster-causing elements, providing an interpretable theoretical framework for anomaly identification. Furthermore, an anomaly index extraction model is constructed based on the knowledge model to effectively capture the abnormal states of disaster-causing elements. This provides a highly reliable risk assessment basis for subsequent spatiotemporal dual-dimensional fire point identification.

[0055] In one feasible implementation, step S01 may include steps S011 to S013: Step S011: Statistical methods are used to extract the temporal dynamic characteristics and spatial differentiation characteristics of the disaster-causing factors and the drought-stricken forest fire areas, respectively. It should be noted that in the embodiments of this application, statistical methods refer to mathematical methods that reveal the quantitative relationships and patterns between variables through the collection, organization, analysis and interpretation of data, including descriptive statistics (mean, variance), inferential statistics (correlation analysis, regression analysis), spatiotemporal statistics (time series decomposition and spatiotemporal autocorrelation analysis), etc.

[0056] Temporal dynamic characteristics refer to the changing patterns of disaster-causing factors or forest fire events over time, such as seasonal fluctuations, daily variation trends, and interannual cycles.

[0057] Spatial differentiation characteristics refer to the different distribution patterns of disaster-causing factors or forest fire events between different locations, such as differences in latitude gradient and altitude, and spatial pattern differences caused by vegetation type differentiation.

[0058] The core objective of this step is to extract the dynamic correlation characteristics between disaster-causing factors and forest fires from the time dimension using statistical methods, and to extract their distribution difference characteristics from the spatial dimension, so as to provide basic data support for the subsequent construction of spatiotemporal distribution relationships.

[0059] The specific implementation approach is as follows: For the extraction of time dynamic features, time series analysis methods are used, such as trend fitting and periodic detection of meteorological factors at daily, monthly, and seasonal scales to identify the abrupt change nodes of disaster-causing factors (such as the time point of a sudden increase in the number of drought days) and the temporal coupling of forest fire events; for the extraction of spatial differentiation features, spatial statistical methods are used, such as analyzing the spatial clustering patterns of disaster-causing factors through the coefficient of variation and spatial autocorrelation index, and combining hotspot analysis to identify spatial hotspot areas with high forest fire incidence and their corresponding factor characteristics.

[0060] By using statistical methods to extract the temporal dynamics and spatial differentiation characteristics of the disaster-causing factors and the drought-stricken forest fire areas, it is possible to quantitatively separate the characteristics of the disaster-causing factors and forest fires in the temporal and spatial dimensions, laying the foundation for the discovery of multi-dimensional correlation patterns.

[0061] Step S012: Construct a spatiotemporal distribution relationship based on the aforementioned temporal dynamic features and spatial differentiation features; Based on the dynamic characteristics of time and the differentiation features of space, the spatiotemporal distribution relationship is constructed. That is, by integrating the dynamic trend of the time dimension and the differentiation pattern of the spatial dimension, the regular patterns of how disaster-causing factors induce forest fires under specific spatiotemporal combinations are revealed.

[0062] The core objective of this step is to couple and analyze the results of temporal dynamic feature extraction and spatial differentiation feature extraction to construct a "spatiotemporal dual-dimensional" correlation model between disaster-causing factors and forest fires, providing a quantitative basis for the knowledge model of drought and forest fire disaster chains.

[0063] Specifically, a four-dimensional database of factors, time, space, and forest fire is established to integrate time series features and spatial differentiation features, and to extract association rules with clear spatiotemporal boundaries. For example, in areas where the number of days of spring drought lasts longer than 10 days and the vegetation cover type is grassland, the probability of forest fire is 50% higher than in other areas. This step can break through the limitations of traditional single-dimensional analysis and more accurately capture the spatiotemporal threshold conditions for the occurrence of disaster chains.

[0064] Step S013: Construct a knowledge model of drought and forest fire disaster chain based on the spatiotemporal distribution relationship.

[0065] To transform spatiotemporal distribution relationships into a computable and interpretable knowledge model, providing rule guidance and theoretical constraints for subsequent anomaly index extraction models, the following steps are taken: First, key patterns in spatiotemporal distribution relationships are extracted and formalized into rules. For example, the statement "the probability of forest fires increases when SPI < -1.5 and vegetation drought index > 0.6" is transformed into an IF-THEN rule (IF condition, THEN conclusion confidence level). Then, the rules are organized using knowledge graphs or causal graphs to clarify the hierarchical relationships between disaster-causing factors (e.g., meteorological factors → vegetation drought → combustible material moisture content → fire). Finally, an uncertainty handling mechanism is introduced, such as adding confidence levels to the rules (e.g., "this rule holds true in 85% of historical cases"), forming a complete knowledge model.

[0066] By constructing knowledge models, complex statistical patterns can be transformed into structured knowledge that machines can understand, thereby improving the interpretability and reliability of subsequent models.

[0067] In one feasible implementation, step S02 may include steps S021 to S023: Step S021: The historical disaster-causing factors in the disaster-causing factors are divided into a training set and a test set using the hold-out method; It should be noted that, in the embodiments of this application, the hold-out method refers to a sample partitioning method that divides the dataset into mutually independent training and test sets according to a certain ratio. During the partitioning process, the randomness and representativeness of the data should be maintained to avoid overfitting or underfitting of the model due to uneven sample distribution. Specifically, the hold-out method is used to divide the historical disaster-causing factors in the historical wildfire event catalog of the collected multi-source dataset into training and test sets.

[0068] Step S022: Select an adaptation algorithm based on the training set and the drought-forest fire disaster chain knowledge model, and train the adaptation algorithm to obtain an initial drought-forest fire disaster chain anomaly index extraction model; The adaptation algorithm refers to the algorithm selected based on the knowledge model rules and the characteristics of the disaster-causing factors, which can effectively learn the mapping relationship between input and output. It must meet the constraints of the knowledge model (such as thresholds and causal relationships in the rules) and the data type of the disaster-causing factors (such as continuous and fractional types). The initial drought-forest fire disaster chain anomaly index extraction model refers to a model initially constructed using training set data and adaptation algorithms. Its parameters are optimized through the training process, but have not yet been validated on a test set, and may have problems such as overfitting or unreasonable parameters.

[0069] The core objective of this step is to select a suitable algorithm based on the guidance of the knowledge model, and to train a model capable of initially outputting anomaly indices using training data, thus transforming knowledge rules into the decision-making logic of the algorithm. The implementation approach is as follows: First, select the appropriate algorithm based on the type of knowledge model. For example, if the knowledge model contains probability rules (such as "the probability of fire increases by 3 times when SPI < -1.5"), Bayesian probability algorithm is preferred; if it contains nonlinear threshold relationships (such as "abnormal risk increases suddenly when vegetation drought index > 0.6"), tree models such as random forest and XGBoost can be selected.

[0070] Furthermore, the disaster-causing factors in the training set are used as input, and the anomaly index (or fire label) is used as output. The model parameters (such as the prior probability of the Bayesian model and the splitting threshold of the tree model) are optimized through algorithm iteration. Finally, the algorithm hyperparameters are adjusted through validation within the training set to obtain an initial model with better performance.

[0071] This step combines the qualitative rules of the knowledge model with data-driven algorithms to achieve preliminary prediction of the anomaly index.

[0072] Step S023: The initial drought-forest fire disaster chain anomaly index extraction model is validated and optimized using the test set to obtain the drought-forest fire disaster chain anomaly index extraction model.

[0073] It should be noted that, in the embodiments of this application, verification optimization refers to evaluating the performance of the initial model through test set data and adjusting the model parameters or structure based on the evaluation results in order to improve the generalization ability and reliability of the model.

[0074] The core purpose of this step is to examine the performance of the initial model using independent test set data, identify and correct problems such as model overfitting and unreasonable thresholds, and ensure that the model can accurately output anomaly indices in practical applications.

[0075] The implementation approach is as follows: First, input the disaster-causing factors of the test set into the initial model to obtain the predicted anomaly index; then, calculate the model performance index by comparing the prediction results with the real labels (such as whether a fire has occurred in the test set); if the index does not reach the preset threshold, analyze the reasons for the error and optimize the model by adjusting the algorithm hyperparameters, supplementing training data, or correcting knowledge rules; finally, retest the optimized model until the performance meets the target and obtain the final model.

[0076] Based on the first and / or second embodiments of this application, a third embodiment of this application is proposed. In the third embodiment of this application, content that is the same as or similar to the first and / or second embodiments described above can be referred to the above description and will not be repeated hereafter.

[0077] Based on this, please refer to Figure 3 , Figure 3 The flowchart provided for the third embodiment of this application is as follows: Figure 3 As shown, step S10, which involves collecting multi-source datasets of drought-stricken forest fire areas and extracting disaster-causing factors from the multi-source data, may include steps S11 to S13: Step S11: Collect multi-source datasets from drought and forest fire areas, and standardize the multi-source datasets to obtain standardized datasets; Standardization refers to the process of unifying the format, correcting biases, and removing noise from collected multi-source data to eliminate differences between different data sources and ensure data consistency and usability. A standardized dataset refers to a collection of data that has undergone standardization, resulting in a unified format, reliable quality, and direct usability for subsequent analysis.

[0078] By systematically collecting various relevant data from drought-stricken forest fire areas and standardizing them, a high-quality data foundation is provided for subsequent extraction of disaster-causing factors and model construction. Integrating multi-source data enables comprehensive monitoring of the drought-stricken forest fire disaster chain, while standardization effectively reduces data errors and improves the accuracy of subsequent analysis results.

[0079] In one possible implementation, the acquisition of multi-source datasets can be achieved in real time through automated data interfaces by acquiring data from meteorological satellites (such as the FY4 satellite), ground meteorological stations, and historical disaster databases; the standardization process can employ machine learning-based adaptive correction algorithms to dynamically eliminate system biases between different satellite sensors.

[0080] Step S12: Extract drought process-related factors from meteorological data in the standardized dataset, and extract drought-forest fire disaster-related factors from satellite data in the standardized dataset; Drought-related factors refer to key indicators extracted from meteorological data that reflect the occurrence and development of drought, such as the Standardized Precipitation Factor (SPI), daily average evaporation, drought duration, and drought level. Among them, the Standardized Precipitation Factor (SPI) is an index obtained from standardized precipitation data to measure the deviation of precipitation from the long-term average in a certain region.

[0081] Drought-related factors in forest fire disasters refer to parameters extracted from satellite data that are directly related to drought-related forest fire disasters, including combustible moisture content and vegetation drought index. Combustible moisture content refers to the proportion of water in combustibles such as vegetation, which is an important factor affecting the occurrence and spread of forest fires; vegetation drought index refers to an index that reflects the degree of drought stress on vegetation (such as the NDVI-Ts spatial drought index) obtained by inversion from satellite remote sensing data.

[0082] The purpose of this step is to accurately extract key factors closely related to drought processes and forest fire disasters from standardized datasets, providing data support for subsequent construction of disaster-causing factors. By separating core indicators from meteorological and satellite data, the driving effect of drought on forest fires and the impact of vegetation and combustible material conditions on disasters can be analyzed in a targeted manner.

[0083] In one possible implementation, the extraction of drought-related factors can be combined with multiple meteorological parameters, such as considering both monthly and seasonal precipitation data when calculating the standardized precipitation factor (SPI).

[0084] Step S13: Generate disaster-causing factors based on the drought process-related factors and the drought-forest fire disaster-related factors, and in conjunction with the historical wildfire event catalog in the standardized dataset.

[0085] The historical wildfire event catalog refers to a dataset that records past wildfire events in the study area, including information such as the time, location, and intensity of the fires, used to correlate causative factors with actual disaster outcomes.

[0086] It should be noted that disaster-causing factors include real-time disaster-causing factors and historical disaster-causing factors. Historical disaster-causing factors refer to disaster-causing factors extracted from multi-source data of actual forest fires in the historical wildfire event catalog. These factors are generated based on historical data and are used for model training and pattern analysis. Real-time disaster-causing factors refer to factors generated based on current monitoring data and used for real-time assessment of disaster risk (such as the current vegetation drought index, real-time temperature, etc.).

[0087] The purpose of this step is to generate a complete system of disaster-causing factors by integrating factors related to drought processes, drought-related forest fire disasters, and a catalog of historical wildfire events. This lays the foundation for subsequent construction of disaster chain knowledge models and anomaly extraction models. By distinguishing between real-time and historical disaster-causing factors, it can meet the needs of real-time monitoring and early warning while also supporting model training based on historical patterns.

[0088] In one possible implementation, the generation of disaster-causing factors can employ a spatiotemporal matching algorithm to associate the fire occurrence time and location in the historical wildfire event catalog with drought-related factors and drought-forest fire disaster-related factors at the corresponding time, forming a historical disaster-causing factor sample; while real-time disaster-causing factors are generated by real-time calculation and dynamic updating of factor data at the current time.

[0089] In this embodiment, by collecting multi-source datasets from drought and forest fire areas and standardizing them, a unified and high-quality data foundation is provided for subsequent analysis, effectively eliminating format differences and noise interference between different data sources. Based on this, drought-related factors and drought-forest fire disaster-related factors are extracted from meteorological and satellite data, respectively, achieving accurate separation and quantification of drought-driving mechanisms and forest fire influencing factors, and improving the targeting and reliability of key indicator extraction. Finally, a disaster-causing factor system containing real-time and historical disaster-causing factors is generated by combining historical wildfire event catalogs, which not only provides dynamic parameter support for real-time risk assessment of drought and forest fire disasters, but also lays a data foundation for training disaster models based on historical patterns.

[0090] Based on the above embodiments of this application, a fourth embodiment of this application is proposed. In this fourth embodiment, content that is the same as or similar to that in the above embodiments can be referred to the above description, and will not be repeated hereafter.

[0091] In this embodiment, step S20, which involves identifying spatiotemporal dual-dimensional thermal anomalies based on the satellite data and combining the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model to obtain the fire point identification result, may include steps S21 to S23: Step S21: Preprocess the satellite data to obtain a standardized satellite brightness temperature dataset; Satellite data refers to Earth surface observation data acquired by sensors carried by spacecraft such as meteorological satellites and remote sensing satellites. This includes, but is not limited to, infrared band data, visible light band data, and thermal infrared band data. It is the main data source for obtaining surface temperature and vegetation information in forest fire monitoring.

[0092] Preprocessing refers to a series of correction, transformation, and optimization operations performed on raw satellite data to eliminate data noise, systematic errors, and format differences, ensuring data quality and usability. A standardized satellite brightness temperature dataset refers to a dataset where the format, units, spatial resolution, and temporal resolution of brightness temperature (or simply "brightness temperature") data have reached a unified standard after preprocessing. Brightness temperature refers to the equivalent blackbody temperature exhibited by an object under thermal radiation, and is an important indicator for measuring the surface temperature of an object.

[0093] The purpose of this step is to systematically preprocess the satellite data, transforming the raw data into a standardized brightness temperature dataset, thus providing a high-quality and consistent data foundation for subsequent thermal anomaly identification. The preprocessing effectively eliminates interference factors such as atmospheric scattering, sensor errors, and terrain shadows, ensuring that the brightness temperature data accurately reflects the true surface temperature characteristics, thereby improving the accuracy and reliability of fire point identification.

[0094] In one possible implementation, preprocessing may include four core steps: radiometric calibration, atmospheric correction, geometric fine correction, and spatial resampling. Radiometric calibration refers to converting the digital quantization values ​​(DN values) output by satellite sensors into physically meaningful radiance values. Atmospheric correction refers to eliminating the influence of atmospheric absorption and scattering on the radiation signal and retrieving the true radiance of the Earth's surface. Geometric fine correction refers to correcting the satellite data to a specified geographic coordinate system using ground control points. Spatial resampling refers to uniformly converting the data to the target spatial resolution to ensure spatial compatibility between different satellite data.

[0095] Step S22: Call the drought and forest fire disaster chain anomaly index extraction model, input the real-time disaster-causing factors into the drought and forest fire disaster chain anomaly index extraction model, and obtain the drought and forest fire disaster chain anomaly index; Real-time disaster-causing factors refer to parameters generated based on current monitoring data that reflect the real-time drought and forest fire risk status, including real-time vegetation drought index, real-time combustible moisture content, and real-time air temperature. The drought-forest fire disaster chain anomaly index refers to a quantitative indicator calculated by a model that comprehensively reflects the probability and intensity of forest fire disaster chains driven by drought; the higher the index value, the higher the anomaly risk of the disaster chain.

[0096] The purpose of this step is to transform real-time disaster-causing factors into drought- and forest fire disaster chain anomaly indices that can be directly used for fire point identification by calling a pre-trained anomaly index extraction model, thereby realizing the dynamic correlation between drought processes and forest fire risks. This index can provide prior knowledge for subsequent thermal anomaly point identification, helping the system distinguish between natural thermal anomalies (such as volcanic activity) and drought-driven forest fire anomalies, thus improving the targeting and accuracy of fire point identification.

[0097] Step S23: Combine the drought forest fire disaster chain anomaly index to perform spatiotemporal dual-dimensional thermal anomaly point identification on the satellite brightness temperature dataset to obtain fire point identification results; wherein, the spatiotemporal dual-dimensional thermal anomaly point identification includes absolute fire point discrimination, conditional fire point discrimination, low temperature fire point monitoring, continuous fire judgment, flare spot removal, and underlying surface judgment.

[0098] Spatiotemporal dual-dimensional thermal anomaly identification refers to a method that comprehensively analyzes satellite brightness temperature data from both temporal and spatial dimensions to identify pixels with significant temperature anomalies (i.e., thermal anomalies). The "temporal dimension" refers to analyzing the brightness temperature variation patterns of the same area at different times, while the "spatial dimension" refers to analyzing the brightness temperature distribution characteristics of different pixels at the same time. Fire point identification results refer to the dataset obtained after thermal anomaly identification, containing information such as fire point location, fire point intensity, and fire point duration; this is the core output of forest fire monitoring.

[0099] The purpose of this step is to combine the drought forest fire disaster chain anomaly index with sub-steps such as absolute fire point discrimination and conditional fire point discrimination to accurately identify real fire points from satellite brightness temperature data, while eliminating false fire points caused by interference factors such as flares and high-temperature underlying surfaces. Spatiotemporal dual-dimensional analysis can fully utilize the time series characteristics and spatial distribution characteristics of brightness temperature data, significantly improving the sensitivity (especially low-temperature fire points) and specificity (reducing false positives) of fire point identification.

[0100] In this embodiment, through systematic data preprocessing, dynamic risk index fusion, and spatiotemporal multi-dimensional fire point screening, accurate identification and comprehensive monitoring of forest fire points driven by drought are achieved. Specifically, standardized preprocessing provides a high-quality brightness and temperature data foundation for fire point identification, effectively eliminating noise such as atmospheric interference and sensor errors. The introduction of a drought-forest fire disaster chain anomaly index dynamically correlates drought and forest fire risks, providing a priori risk background for fire point identification. This significantly improves the accuracy, sensitivity, and scene adaptability of fire point identification, providing reliable technical support for early forest fire warnings and emergency responses.

[0101] In one feasible implementation, step S23 may include steps S231 to S236: Step S231: Combine the drought forest fire disaster chain anomaly index to perform absolute fire point discrimination on the satellite brightness temperature dataset to obtain preliminary absolute fire point identification results; Absolute fire point discrimination refers to the method of directly filtering out pixels whose brightness temperature exceeds a preset brightness temperature threshold and initially identifying them as potential fire points. It is the basic screening step for fire point identification.

[0102] The preliminary identification result of absolute fire points refers to the set of pixels that only contain brightness temperature values ​​exceeding the threshold obtained after absolute fire point discrimination. It has not been verified by other conditions and may contain false fire points.

[0103] The purpose of this step is to dynamically adjust the brightness temperature threshold using the drought-fire disaster chain anomaly index, quickly screening out potential fire points with significant high-temperature characteristics and reducing the amount of data required for subsequent processing. By introducing the anomaly index, the adaptability problem of fixed thresholds in different seasons and regions can be avoided. For example, the threshold can be lowered in areas with high drought risk to capture early fire points, while the threshold can be increased in humid areas to reduce false positives.

[0104] Step S232: Perform conditional fire point discrimination on the satellite brightness temperature dataset from the time dimension and the spectral dimension to obtain preliminary conditional fire point identification results; the preliminary conditional fire point identification results include conditional fire point pixels and brightness temperature features corresponding to the conditional fire point pixels; The time dimension refers to analyzing the brightness temperature change trend of the same pixel at multiple consecutive moments (such as the rate of temperature increase within 1 hour) to identify short-term abnormal temperature rise signals; the spectral dimension refers to analyzing the brightness temperature difference of the same pixel in different spectral bands and using the difference in radiation characteristics of the fire point in different bands for discrimination.

[0105] Conditional fire point discrimination refers to a method that combines temporal and spectral features to screen pixels that do not reach the absolute fire point threshold but exhibit abnormal heating characteristics. Conditional fire point pixels are those that, through conditional fire point discrimination, possess potential fire point characteristics but whose brightness temperature does not reach the absolute threshold; brightness temperature characteristics refer to the ratio of the heating rate (e.g., 5K per hour) of the conditional fire point pixel in the temporal dimension to the band ratio in the spectral dimension.

[0106] The purpose of this step is to overcome the limitations of absolute fire point discrimination by capturing early low-temperature fire points or weak fire points (such as smoldering fires) through time and spectral characteristics, avoiding missed detections due to brightness temperatures not reaching the absolute threshold. Conditional fire point discrimination can utilize the dynamic changes of fire points over time and the specificity of spectral characteristics to help the system distinguish between natural thermal anomalies (such as solar radiation) and drought-driven forest fire anomalies.

[0107] Step S233: Capture the tiny temperature rise signal of the conditional fire point pixel to perform low-temperature fire point monitoring, and monitor the conditional fire point pixel from the time dimension to determine continuous fire. A minute temperature rise signal refers to a slight temperature increase that occurs within a short period of time in a conditional ignition pixel. This signal may indicate an early smoldering fire or a low-intensity ignition, such as the burning of dead branches and leaves on the ground. Low-temperature ignition monitoring refers to capturing minute temperature rise signals using a high-sensitivity temperature detection algorithm to identify ignitions whose temperature is below the absolute ignition threshold but shows a continuous temperature rise trend.

[0108] Continuous fire detection refers to monitoring the duration of brightness temperature changes in conditional fire pixels over time to determine whether they have formed a continuous combustion state (e.g., fire pixels persist for more than 1 hour and brightness temperature fluctuation is less than 5%).

[0109] The purpose of this step is to further improve the sensitivity and timeliness of fire point identification, to achieve early detection of low-temperature fire points by capturing minute temperature rise signals, and to distinguish between transient thermal anomalies (such as heat radiation from vehicle exhaust) and real, continuous fire points by judging continuous fires, thereby reducing misjudgments caused by instantaneous interference.

[0110] Step S234: The preliminary identification result of the absolute fire point is merged and deduplicated with the conditional fire point pixels obtained by the low temperature fire point monitoring and the continuous fire judgment to obtain a fire point pixel set. Merging and deduplication refers to the process of merging two sets of cells into one dataset and removing cells that have duplicate spatial locations (i.e., only one cell from the same geographic location is retained).

[0111] The fire point pixel set refers to the dataset containing all potential fire point pixels after merging and deduplication, and it serves as the input data for subsequent flare removal and underlying surface verification.

[0112] The purpose of this step is to integrate two types of potential fire sources: absolute fire points and conditional fire points, to form a complete set of fire point pixels. This ensures that both high-temperature fire points (such as open flames) and low-temperature fire points (such as smoldering fires) are included in subsequent analysis, avoiding the omission of fire points due to a single source. The merging and deduplication operation can eliminate duplicate pixels (such as the same fire point being captured by both absolute and conditional fire point discrimination), reducing the amount of data for subsequent processing.

[0113] In one possible implementation, deduplication can be achieved through spatial location matching: using the latitude and longitude coordinates of a pixel as a unique identifier, when the latitude and longitude deviation between the absolute fire point pixel and the conditional fire point pixel is less than 0.5 pixels, it is determined to be a duplicate pixel, and only the pixel with the higher brightness temperature value is retained; for non-duplicate pixels, all are retained and integrated into a fire point pixel set.

[0114] Step S235: Remove flare spots from the fire point pixel set to obtain the fire point candidate set; It should be noted that, in the embodiments of this application, flares refer to bright pixels generated by satellite sensors during shooting due to direct sunlight, reflection from water or metal surfaces, etc. Their spectral characteristics are similar to fire spots (such as high brightness temperature values), but they are not real fire spots and are one of the main interference factors in fire spot identification.

[0115] Fire candidate set refers to the set of real fire candidate pixels that have been preliminarily screened after removing flares.

[0116] The purpose of this step is to identify and remove flares from the fire point pixel set through joint discrimination of multi-band data, thereby reducing the impact of non-fire point interference factors on the identification results. Flares typically have significant spectral characteristics (such as extremely high reflectivity in the visible light band and no abnormal brightness temperature in the thermal infrared band) and temporal characteristics (such as instantaneous appearance and short duration), which can be effectively distinguished through multi-dimensional features.

[0117] Step S236: Verify the underlying surface type of the candidate fire point set to obtain the fire point identification result.

[0118] Underlying surface type verification refers to the process of determining whether a pixel is a potential forest fire area (such as forest or grassland) by analyzing the land cover type corresponding to the pixel. Underlying surface type refers to the cover features of the Earth's surface, including types such as forests, grasslands, farmland, water bodies, cities, and bare land; Fire point identification results refer to the final confirmed dataset of real forest fire points after verification of underlying surface type, which includes information such as the geographical location (latitude and longitude), occurrence time, fire intensity (brightness temperature value), and duration of the fire points.

[0119] The purpose of this step is to further filter the candidate set of fire points by combining geographical background information, so as to ensure that the identification results only include fire points located in areas where forest fires may occur (such as arid forest areas), and exclude non-forest fire-driven thermal anomalies such as urban high temperatures, industrial heat sources, and volcanoes, thereby improving the accuracy and application value of fire point identification.

[0120] In one possible implementation, underlying surface type verification can be achieved through spatial overlay of land use / cover data: the geographic location of the candidate fire point pixels is spatially overlaid with high-resolution cover data, and when the underlying surface type corresponding to the pixel is "forest", "grassland" or "shrubland", the pixel is retained; if it is "city", "water body", "farmland", "bare land" or "glacier", it is determined to be a non-forest fire point and removed.

[0121] For example, to help understand the implementation process of the drought and forest fire hazard extraction and identification method obtained by combining the above embodiments, please refer to... Figure 4 , Figure 4A simplified flowchart illustrating a method for extracting and identifying drought-induced forest fire disaster factors is provided, specifically: Step 1. Data Collection and Preprocessing: We collected a catalog of historical wildfire events, meteorological data, and satellite data for the study area. Satellite data was used to extract data on combustible moisture content and vegetation drought index related to the drought-fire disaster chain; meteorological data was used to extract data factors related to the drought process, such as the standardized precipitation factor (SPI), daily average evaporation, drought duration, and drought level. The collected data underwent preprocessing, including data correction, noise reduction, and format conversion.

[0122] Step 2. Study on the spatiotemporal distribution patterns of disaster-causing factors and forest fires Based on traditional statistics, this study investigates the spatiotemporal distribution relationship between pre-processed disaster-causing factors and forest fires, obtains preliminary spatiotemporal distribution patterns of disaster-causing elements and fires, and constructs a knowledge model of the occurrence and development process of drought forest fire disaster chains, providing a theoretical basis for subsequent abnormal feature extraction.

[0123] Step 3. Construction of an anomaly index extraction model for drought and forest fire disaster chains Building upon step 2, a knowledge-driven anomaly extraction model for disaster-causing factors in the drought-forest fire disaster chain is constructed based on Bayesian probability, deep learning, and other technologies. The specific steps are as follows: Dataset partitioning: The dataset is partitioned using the hold-out method, with 80% used for training and 20% for testing.

[0124] Algorithm selection and application: Anomaly extraction research was conducted using algorithms such as Bayesian probability, random forest, decision tree, and XGBoost.

[0125] Model validation: Validate the constructed model to ensure its accuracy and reliability.

[0126] Step 4. Identification of thermal anomalies in a spatiotemporal dual-dimensional combination A comprehensive judgment is made based on a combination of spatiotemporal dimensions. This involves absolute fire point discrimination, conditional fire point discrimination, monitoring the high-temperature threshold of pixels, temperature changes in pixel time series, and differences between neighboring pixels to comprehensively monitor fire points. The specific method is as follows: Absolute fire point identification: Taking FY4 data as an example, when the brightness temperature value of the B7 band is higher than 360K, it is directly identified as a forest-grassland fire point.

[0127] Conditional fire identification: Based on the fact that the temperature of ground objects can only change by 1K within 1015 minutes under normal conditions, when the temperature change of ground objects reaches more than 5K within 1015 minutes, and the brightness temperature difference between the B7 band and the B12 band is 10K (20K during the day), a forest fire is considered to have occurred.

[0128] Low-temperature fire spot monitoring: By utilizing the brightness temperature change between the previous pixel value and the current pixel value, low-temperature fire spots can be effectively monitored and small forest fires can be detected.

[0129] Continuous fire assessment: When continuous fires are detected occurring in the same location using time-based monitoring, the current forest fire status is determined to be continuous forest fire.

[0130] Flare removal: Conditional fire point discrimination and flare removal are performed based on the discriminant formula.

[0131] Underlying surface assessment: Using 30-meter ground cover data, determine whether the underlying surface is forest-steppe, and exclude fire points in non-forest-steppe areas.

[0132] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the method for extracting and identifying drought and forest fire disaster-causing elements in this application. Any simple modifications based on this technical concept are within the scope of protection of this application.

[0133] This application also provides a device for extracting and identifying drought and forest fire disaster-causing factors. Please refer to [reference needed]. Figure 5 The drought and forest fire hazard extraction and identification device includes: Data collection module 10 is used to collect multi-source datasets of drought and forest fire areas and extract disaster-causing factors from the multi-source datasets, which include satellite data; The thermal anomaly identification module 20 is used to identify thermal anomalies in a spatiotemporal dual dimension based on the satellite data, and to obtain the fire point identification result by combining the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model; wherein, the drought and forest fire disaster chain anomaly index extraction model is constructed based on the drought and forest fire disaster chain knowledge model, and the drought and forest fire disaster chain knowledge model is obtained by analyzing the spatiotemporal distribution relationship of the disaster-causing factors.

[0134] The drought- and forest fire-causing element extraction and identification device provided in this application, employing the drought- and forest fire-causing element extraction and identification method described in the above embodiments, can solve the technical problem of drought- and forest fire-causing element extraction and identification. Compared with the prior art, the beneficial effects of the drought- and forest fire-causing element extraction and identification device provided in this application are the same as those of the drought- and forest fire-causing element extraction and identification method described in the above embodiments, and other technical features in the drought- and forest fire-causing element extraction and identification device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0135] This application provides a device for extracting and identifying drought and forest fire disaster-causing factors. The device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the drought and forest fire disaster-causing factor extraction and identification method in the above embodiment 1.

[0136] The following is for reference. Figure 6 The diagram illustrates a structural schematic of a drought and forest fire hazard extraction and identification device suitable for implementing embodiments of this application. The drought and forest fire hazard extraction and identification device in this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and vehicle terminals (e.g., vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 6 The drought and forest fire disaster-causing element extraction and identification device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0137] like Figure 6As shown, the drought and forest fire hazard extraction and identification device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the drought and forest fire hazard extraction and identification device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, a touch screen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. The communication device 1009 allows the drought and forest fire hazard extraction and identification equipment to communicate wirelessly or wiredly with other equipment to exchange data. Although the figure shows drought and forest fire hazard extraction and identification equipment with various systems, it should be understood that it is not required to implement or possess all of the systems shown. More or fewer systems may be implemented alternatively.

[0138] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a 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, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0139] The drought- and forest fire-causing element extraction and identification device provided in this application, employing the drought- and forest fire-causing element extraction and identification method described in the above embodiments, can solve the technical problem of drought- and forest fire-causing element extraction and identification. Compared with the prior art, the beneficial effects of the drought- and forest fire-causing element extraction and identification device provided in this application are the same as those of the drought- and forest fire-causing element extraction and identification method described in the above embodiments, and other technical features of the device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0140] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0141] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0142] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the drought and forest fire disaster-causing element extraction and identification method in the above embodiments.

[0143] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0144] The aforementioned computer-readable storage medium may be included in the drought and forest fire hazard extraction and identification device; or it may exist independently and not be assembled into the drought and forest fire hazard extraction and identification device.

[0145] The aforementioned computer-readable storage medium carries one or more programs. When the aforementioned one or more programs are executed by the drought and forest fire disaster-causing element extraction and identification device, the drought and forest fire disaster-causing element extraction and identification device: collects a multi-source dataset of drought and forest fire areas and extracts disaster-causing factors from the multi-source dataset, wherein the multi-source dataset includes satellite data. Based on the satellite data, spatiotemporal dual-dimensional thermal anomaly identification is performed, and combined with the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model, the fire point identification result is obtained; wherein, the drought and forest fire disaster chain anomaly index extraction model is constructed based on the drought and forest fire disaster chain knowledge model, which is obtained based on the spatiotemporal distribution relationship analysis of the disaster-causing factors.

[0146] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0147] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, 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 indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated 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 the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0148] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0149] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described method for extracting and identifying drought and forest fire disaster-causing elements, thereby solving the technical problem of extracting and identifying drought and forest fire disaster-causing elements. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the drought and forest fire disaster-causing element extraction and identification method provided in the above embodiments, and will not be repeated here.

[0150] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the drought and forest fire disaster-causing element extraction and identification method described above.

[0151] The computer program product provided in this application can solve the technical problem of extracting and identifying drought and forest fire disaster-causing factors. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the drought and forest fire disaster-causing factor extraction and identification method provided in the above embodiments, and will not be repeated here.

[0152] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for extracting and identifying drought-induced forest fire disaster factors, characterized in that, The method for extracting and identifying drought and forest fire causative factors includes: Collect multi-source datasets of drought and forest fire areas, and extract disaster-causing factors from the multi-source datasets, which include satellite data; Based on the satellite data, spatiotemporal dual-dimensional thermal anomaly identification is performed, and combined with the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model, the fire point identification result is obtained; wherein, the drought and forest fire disaster chain anomaly index extraction model is constructed based on the drought and forest fire disaster chain knowledge model, which is obtained based on the spatiotemporal distribution relationship analysis of the disaster-causing factors.

2. The method for extracting and identifying drought and forest fire causative factors as described in claim 1, characterized in that, Before the step of performing spatiotemporal dual-dimensional thermal anomaly identification based on the satellite data and combining it with the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model to obtain the fire point identification result, the following steps are also included: Analyze the spatiotemporal distribution relationship between the disaster-causing factors and the drought-stricken forest fire areas, and construct a drought-stricken forest fire disaster chain knowledge model based on the spatiotemporal distribution relationship; Based on the drought and forest fire disaster chain knowledge model and the disaster-causing factors, an abnormal index extraction model for the drought and forest fire disaster chain is constructed.

3. The method for extracting and identifying drought and forest fire disaster-causing factors as described in claim 2, characterized in that, The steps of analyzing the spatiotemporal distribution relationship between the disaster-causing factors and the drought-stricken forest fire areas, and constructing a drought-stricken forest fire disaster chain knowledge model based on the spatiotemporal relationship, include: Statistical methods were used to extract the temporal dynamics and spatial differentiation characteristics of the disaster-causing factors and the drought-stricken forest fire areas, respectively. A spatiotemporal distribution relationship is constructed based on the aforementioned temporal dynamic characteristics and spatial differentiation characteristics; A knowledge model of drought and forest fire disaster chain is constructed based on the spatiotemporal distribution relationship.

4. The method for extracting and identifying drought- and forest fire-causing factors as described in claim 2, characterized in that, The steps of constructing the drought and forest fire disaster chain anomaly index extraction model based on the drought and forest fire disaster chain knowledge model and the disaster-causing factors include: The historical disaster-causing factors among the disaster-causing factors are divided into training and testing sets using the hold-out method. Based on the training set and the drought and forest fire disaster chain knowledge model, an adaptation algorithm is selected and trained to obtain an initial drought and forest fire disaster chain anomaly index extraction model. The initial drought-forest fire disaster chain anomaly index extraction model was validated and optimized using the test set, resulting in the drought-forest fire disaster chain anomaly index extraction model.

5. The method for extracting and identifying drought- and forest fire-causing factors as described in claim 1, characterized in that, The steps of collecting multi-source datasets of drought and forest fire areas and extracting disaster-causing factors from the multi-source data include: Multi-source datasets from arid forest fire areas were collected, and the multi-source datasets were standardized to obtain standardized datasets. Factors related to drought processes are extracted from meteorological data in the standardized dataset, and factors related to drought and forest fire disasters are extracted from satellite data in the standardized dataset. Disaster-causing factors are generated based on the drought process-related factors and the drought-forest fire disaster-related factors, combined with the historical wildfire event catalog in the standardized dataset.

6. The method for extracting and identifying drought and forest fire disaster-causing factors as described in claim 1, characterized in that, The steps for identifying fire points based on the aforementioned satellite data in a spatiotemporal dual-dimensional thermal anomaly, and combining this with the drought and forest fire disaster chain anomaly index output by a pre-constructed drought and forest fire disaster chain anomaly index extraction model, include: The satellite data is preprocessed to obtain a standardized satellite brightness temperature dataset; The drought and forest fire disaster chain anomaly index extraction model is invoked, and the real-time disaster-causing factors among the disaster-causing factors are input into the drought and forest fire disaster chain anomaly index extraction model to obtain the drought and forest fire disaster chain anomaly index. Based on the drought and forest fire disaster chain anomaly index, the satellite brightness temperature dataset is used to identify thermal anomalies in a spatiotemporal dimension to obtain fire point identification results. The spatiotemporal dual-dimensional thermal anomaly identification includes absolute fire point discrimination, conditional fire point discrimination, low temperature fire point monitoring, continuous fire judgment, flare spot removal, and underlying surface judgment.

7. The method for extracting and identifying drought and forest fire causative factors as described in claim 6, characterized in that, The step of combining the drought and forest fire disaster chain anomaly index to perform spatiotemporal dual-dimensional thermal anomaly point identification on the satellite brightness temperature dataset to obtain fire point identification results includes: By combining the drought forest fire disaster chain anomaly index with the satellite brightness temperature dataset, absolute fire point identification is performed to obtain preliminary absolute fire point identification results; Conditional fire point discrimination is performed on the satellite brightness temperature dataset from both temporal and spectral dimensions to obtain preliminary conditional fire point identification results; the preliminary conditional fire point identification results include conditional fire point pixels and brightness temperature features corresponding to the conditional fire point pixels. The system captures minute temperature rise signals of the conditional fire point pixels to perform low-temperature fire point monitoring, and monitors the conditional fire point pixels from a time dimension to determine continuous fires. The preliminary identification results of absolute fire points are merged and deduplicated with the conditional fire point pixels obtained by the low-temperature fire point monitoring and the continuous fire judgment to obtain a fire point pixel set. Flare spots are removed from the fire point pixel set to obtain the fire point candidate set; The underlying surface type is verified on the candidate fire point set to obtain the fire point identification result.

8. A system for extracting and identifying drought-induced forest fire disaster factors, characterized in that, The drought and forest fire causation factor extraction and identification system includes: The data collection module is used to collect multi-source datasets of drought and forest fire areas and extract disaster-causing factors from the multi-source datasets, which include satellite data. The thermal anomaly identification module is used to identify thermal anomalies in a spatiotemporal dual dimension based on the satellite data, and to obtain the fire point identification result by combining the drought and forest fire disaster chain anomaly index output by the pre-constructed drought and forest fire disaster chain anomaly index extraction model; wherein, the drought and forest fire disaster chain anomaly index extraction model is constructed based on the drought and forest fire disaster chain knowledge model, which is obtained by analyzing the spatiotemporal distribution relationship of the disaster-causing factors.

9. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the drought and forest fire disaster-causing element extraction and identification method as described in any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the drought and forest fire disaster-causing element extraction and identification method as described in any one of claims 1 to 7.

Images