Locust suitable growing area monitoring method, device and equipment, storage medium and program product

Abstract

The invention provides a locust suitable growing area monitoring method, device and equipment, a storage medium and a program product, and relates to the technical field of artificial intelligence, and the method comprises the steps: obtaining a locust suitable growing area monitoring index set and a locust host suitable growing area monitoring index set; the locust suitable region monitoring index set is determined based on environmental factors affecting growth and development of locusts, the locust host suitable region monitoring index set is determined based on environmental factors affecting distribution of locust hosts, and the environmental factors include climate factors, vegetation factors, land utilization type factors and terrain factors; and inputting the locust suitable growing area monitoring index set and the locust host suitable growing area monitoring index set into a suitable growing area monitoring model to obtain locust suitable growing area monitoring results of the target area in a future preset time period. Through the mode, monitoring errors caused by single monitoring indexes can be avoided, the accuracy of the monitoring result of the locust suitable growing area can be effectively improved, and accurate dynamic monitoring of the locust suitable growing area in the future is achieved.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, apparatus, equipment, storage medium, and program product for monitoring suitable locust habitats. Background Technology

[0002] Grasslands are an important component of terrestrial ecosystems, playing a vital role in water conservation, soil and water retention, and maintaining biodiversity. When locust plagues occur on grasslands, the grassland ecosystem is damaged; therefore, locust control is crucial.

[0003] Because individual locusts are small, it is difficult to identify them directly from remote sensing images. Moreover, the occurrence and development of locust plagues are not random, but are regulated by complex mechanisms of interaction between the locust source, host, and environment. Therefore, in order to meet the needs of large-scale locust monitoring, some scholars no longer choose to directly identify single locust images, but instead choose to use multi-source remote sensing data to dynamically monitor locust habitats, monitor the future changing trends of locust habitats in specific areas, and thus provide support for future locust plague prevention and early warning.

[0004] Some related technologies utilize meteorological data from specific regions to monitor future trends in suitable locust habitats. However, changes in suitable locust habitats are not solely influenced by climatic factors, and relying solely on meteorological data to obtain trends is inaccurate.

[0005] Therefore, existing methods for monitoring suitable locust habitats suffer from low accuracy, making it difficult to achieve accurate monitoring of suitable locust habitats. Summary of the Invention

[0006] This invention provides a method, apparatus, equipment, storage medium, and program product for monitoring suitable locust habitats, which solves the problem of low accuracy in existing methods for monitoring suitable locust habitats and makes it difficult to achieve accurate monitoring of suitable locust habitats.

[0007] This invention provides a method for monitoring suitable locust habitats, comprising: acquiring a set of monitoring indicators for suitable locust habitats and a set of monitoring indicators for suitable locust host habitats in a target area; the set of monitoring indicators for suitable locust habitats is determined based on environmental factors affecting locust growth and development, and the set of monitoring indicators for suitable locust host habitats is determined based on environmental factors affecting the distribution of locust hosts, including climate factors, vegetation factors, land use type factors, and topographic factors; inputting the set of monitoring indicators for suitable locust habitats and the set of monitoring indicators for suitable locust host habitats into a pre-trained suitable habitat monitoring model to obtain the monitoring results of suitable locust habitats in the target area within a future preset time period output by the suitable habitat monitoring model; wherein, the suitable habitat monitoring model is trained based on a sample set of monitoring indicators for suitable locust habitats and a sample set of monitoring indicators for suitable locust host habitats in a sample area.

[0008] According to the present invention, a method for monitoring suitable locust habitats is provided. The suitable habitat monitoring model is a heterogeneous ensemble model constructed based on the RF algorithm, XGBoost algorithm, and LightGBM algorithm. The monitoring results of suitable locust habitats are generated by the suitable habitat monitoring model based on the following steps: Based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set, the suitable habitat is monitored using the RF algorithm to generate a first suitable habitat monitoring result; based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set, the suitable habitat is monitored using the XGBoost algorithm to generate a second suitable habitat monitoring result; based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set, the suitable habitat is monitored using the LightGBM algorithm to generate a third suitable habitat monitoring result; based on the first suitable habitat monitoring result, the second suitable habitat monitoring result, and the third suitable habitat monitoring result, the mean is calculated to generate the final locust suitable habitat monitoring result.

[0009] According to the present invention, a method for monitoring suitable locust habitats includes a set of monitoring indicators for suitable locust habitats, comprising at least one first monitoring indicator affecting locust growth and development. The set of monitoring indicators for suitable locust habitats is determined based on the following steps: obtaining multiple initial first monitoring indicators; performing a difference test on the multiple initial first monitoring indicators to obtain a first difference test result, and selecting at least two candidate first monitoring indicators from the multiple initial first monitoring indicators based on the first difference test result; the candidate first monitoring indicators are monitoring indicators whose influence on locust growth and development is greater than or equal to a first preset threshold; performing a correlation analysis on all candidate first monitoring indicators to obtain a first correlation analysis result, and selecting at least one first monitoring indicator from all candidate first monitoring indicators based on the first correlation analysis result, and adding it to the set of monitoring indicators for suitable locust habitats.

[0010] According to the locust suitable habitat monitoring method provided by the present invention, the first monitoring indicator in the locust suitable habitat monitoring index set includes at least one of the following in the target area: egg stage precipitation, locust nymph stage precipitation, temperature, altitude, slope, isotherm, temperature seasonal variation coefficient, precipitation variation coefficient, locust pre-hatching precipitation, temperature and precipitation during locust oviposition period, and land use type.

[0011] According to the present invention, a method for monitoring suitable locust habitats includes a set of monitoring indicators for suitable locust host habitats, comprising at least one second monitoring indicator that affects the distribution of locust hosts. The set of monitoring indicators for suitable locust host habitats is determined based on the following steps: obtaining multiple initial second monitoring indicators; performing a difference test on the multiple initial second monitoring indicators to obtain a second difference test result, and selecting at least two candidate second monitoring indicators from the multiple initial second monitoring indicators based on the second difference test result; the candidate second monitoring indicators are monitoring indicators whose influence on the distribution of locust hosts is greater than or equal to a second preset threshold; performing a correlation analysis on all candidate second monitoring indicators to obtain a second correlation analysis result, and selecting at least one second monitoring indicator from all candidate second monitoring indicators based on the second correlation analysis result, and adding it to the set of monitoring indicators for suitable locust host habitats.

[0012] According to the locust suitable habitat monitoring method provided by the present invention, the second monitoring index in the locust host suitable habitat monitoring index set includes at least one of the following: annual average temperature, average daily temperature range, isotherm, temperature seasonality coefficient, annual average precipitation, precipitation variation coefficient, total precipitation, locust host growth period precipitation, altitude, slope, aspect, and land use type.

[0013] The present invention also provides a locust suitable habitat monitoring device, comprising: an acquisition module for acquiring a locust suitable habitat monitoring index set and a locust host suitable habitat monitoring index set for a target area; the locust suitable habitat monitoring index set is determined based on environmental factors affecting locust growth and development, and the locust host suitable habitat monitoring index set is determined based on environmental factors affecting locust host distribution, including climate factors, vegetation factors, land use type factors, and topographic factors; and a suitable habitat monitoring module for inputting the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set into a pre-trained suitable habitat monitoring model to obtain the locust suitable habitat monitoring results for the target area within a future preset time period output by the suitable habitat monitoring model; wherein, the suitable habitat monitoring model is trained based on a sample locust suitable habitat monitoring index set and a sample locust host suitable habitat monitoring index set for a sample area.

[0014] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the locust habitat monitoring methods described above.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the locust habitat monitoring methods described above.

[0016] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements any of the locust habitat monitoring methods described above.

[0017] The locust suitable habitat monitoring method, device, equipment, storage medium, and program products provided by this invention no longer rely solely on meteorological data for suitable habitat monitoring. Instead, they fully consider multiple environmental factors affecting locust growth and development. A set of locust suitable habitat monitoring indicators for the target area is constructed based on these factors. Furthermore, considering that changes in locust suitable habitats are also influenced by the distribution of locust hosts, a set of locust host suitable habitat monitoring indicators for the target area is also constructed based on these factors. These indicators are then input into a pre-trained suitable habitat monitoring model to obtain the locust suitable habitat monitoring results for the target area within a preset future time period. This avoids monitoring errors caused by a single indicator, effectively improves the accuracy of locust suitable habitat monitoring results, and enables precise dynamic monitoring of future locust suitable habitats. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the locust habitat monitoring method provided by the present invention.

[0020] Figure 2 This is one of the schematic diagrams of the difference test provided by the present invention.

[0021] Figure 3 This is one of the schematic diagrams of correlation analysis provided by the present invention.

[0022] Figure 4 This is one of the schematic diagrams of the importance test provided by the present invention.

[0023] Figure 5 This is the second schematic diagram of the difference test provided by the present invention.

[0024] Figure 6 This is the second schematic diagram of the correlation analysis provided by the present invention.

[0025] Figure 7 This is the second schematic diagram of the importance test provided by the present invention.

[0026] Figure 8 This is a schematic diagram of the locust habitat monitoring device provided by the present invention.

[0027] Figure 9 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0029] Please see Figures 1 to 7 , Figure 1 This is a flowchart illustrating the locust habitat monitoring method provided by the present invention. Figure 2 This is one of the schematic diagrams of the difference test provided by the present invention. Figure 3 This is one of the schematic diagrams of correlation analysis provided by the present invention. Figure 4 This is one of the schematic diagrams of the importance test provided by the present invention. Figure 5 This is the second schematic diagram of the difference test provided by the present invention. Figure 6 This is the second schematic diagram of the correlation analysis provided by the present invention. Figure 7 This is the second schematic diagram of the importance test provided by the present invention.

[0030] like Figure 1 As shown, in this embodiment, the locust habitat monitoring method includes steps S110 to S120, and the specific steps are as follows: S110: Obtain the monitoring index set of suitable locust habitats and the monitoring index set of suitable locust host habitats in the target area.

[0031] The monitoring index set for suitable locust habitats is determined based on environmental factors that affect locust growth and development, while the monitoring index set for suitable locust host habitats is determined based on environmental factors that affect the distribution of locust hosts. These environmental factors include climate factors, vegetation factors, land use type factors, and topographic factors.

[0032] The target area is the area to be monitored.

[0033] Specifically, based on the growth and development mechanism of grassland locusts, we first explore the response relationship between grassland locusts and different environmental factors, construct a monitoring index system for grassland locust suitable habitats in the target area under future climate scenarios, and generate a monitoring index set for locust suitable habitats in the target area.

[0034] The monitoring index set for suitable locust habitats includes at least one primary monitoring index that affects locust growth and development.

[0035] Generally, the life cycle of grassland locusts can be divided into three stages: the locust egg stage, the locust nymph stage, and the adult stage. The growth and development of grassland locusts at each stage are affected by different environmental factors.

[0036] During the locust egg stage, the locust eggs develop in the soil, mainly influenced by factors such as the minimum temperature, precipitation, soil sand content, and soil pH (potential of hydrogen). Among these, the minimum temperature directly affects the survival rate of locust eggs; excessively low temperatures may cause frost damage or diapause. Precipitation is also a key environmental factor affecting locust eggs; excessive rainfall can easily lead to mold growth on the eggs, while insufficient rainfall can easily cause them to dehydrate and die.

[0037] During the nymph stage, the main influencing factors for grassland locusts include vegetation type, vegetation cover, precipitation, temperature, and the elevation, slope, and aspect of the target area. Among these, vegetation provides habitat and food sources for locust nymphs; precipitation indirectly affects nymph feeding by influencing vegetation growth; the temperature (NT) of the target area affects the developmental process of locust nymphs, as they can only enter the next developmental stage after reaching the initial developmental temperature and accumulating a certain effective accumulated temperature; topographic factors, including the elevation, slope, and aspect of the target area, significantly affect the distribution of locust nymphs. For example, the Asian small locust is mostly distributed in areas with an elevation of 1000 to 2000 meters, while the narrow-winged locust is commonly found in high-altitude areas with an elevation of 4000 to 5000 meters.

[0038] During the adult stage, factors such as temperature, precipitation, vegetation type, and altitude in the target area continue to influence the behavior and reproduction of locust adults. For example, temperature and precipitation (SP) during the locust oviposition period are very important for locust reproduction. Most locust adults choose to lay eggs in sunny weather with temperatures between 26 and 28 degrees Celsius, and low-lying, sunny, concave slopes are often the preferred oviposition sites for locust adults.

[0039] In addition, the seasonal variation coefficient of temperature (TS) and the coefficient of variation of precipitation (PS) in the target area have a significant impact on the growth and development of grassland locusts at all stages.

[0040] Therefore, based on the above analysis, the first monitoring indicator in the locust suitable habitat monitoring index set may include at least one of the following in the target area: egg stage precipitation, locust nymph stage precipitation, temperature, altitude, slope, isotherm, temperature seasonal variation coefficient, precipitation variation coefficient, pre-hatching precipitation of locusts, temperature and precipitation during the locust oviposition period, and land use type.

[0041] Similarly, by combining the growth and reproduction characteristics of grassland vegetation and utilizing future meteorological data, land use data, and topographic data, a monitoring index system for suitable habitats of grassland locusts in the target area under future climate scenarios is constructed, and a monitoring index set for suitable habitats of locusts in the target area is generated.

[0042] Among them, the monitoring index set for suitable locust host areas includes at least one second monitoring index that affects the distribution of locust hosts.

[0043] Grassland vegetation serves as the host for grassland locusts, providing them with habitat and food. It plays a crucial role in the growth, development, and reproduction of grassland locusts. For example, during molting, grassland locusts typically grasp the roots and leaves of grasses and struggle to break free, with the host providing a suitable location. When laying eggs, grassland locusts usually choose depressions with 20% to 50% vegetation cover. The host provides ample sunlight for egg-laying and ensures a sufficient food source, enabling the grassland locusts to successfully complete their egg-laying activities.

[0044] Because different grassland locusts vary in distribution range and growth period, different vegetation types also contribute differently to grassland locust populations. For example, the hairy-footed locust usually causes serious damage to the early growth of grasses, especially sheepgrass, wheatgrass, and artemisia; the Asian small locust usually feeds on grasses and sedges, such as needlegrass, Baikal needlegrass, sedge, and crested wheatgrass; the narrow-winged locust mainly feeds on purple needlegrass, Poa breviscapus-Poa annua, and Altai asters; and the white-edged locust mainly feeds on artemisia, sheepgrass, needlegrass, alkali grass, and leek grass.

[0045] This embodiment comprehensively analyzes the preferred vegetation types of different grassland locusts, selects multiple preferred vegetation types for subsequent analysis, and determines the distribution of various preferred vegetation types within the target area based on vegetation type data. Furthermore, in areas where host vegetation is distributed, multiple distribution points of preferred grassland locust vegetation are determined through random point selection. In areas where non-preferred vegetation is distributed, multiple non-host vegetation distribution points are pseudo-randomly generated for the construction of subsequent monitoring index sets for locust host suitable habitats.

[0046] Generally, the phenological processes of grassland vegetation are mainly divided into the greening-up period, the growing period, and the withering period. Studies have shown that temperature and precipitation are key environmental factors restricting the growth of locust hosts, and the two are interconnected and synergistic. Annual average temperature can affect the growing season and heat accumulation of grassland vegetation; the daily average temperature range has a significant impact on the respiration and photosynthesis of grassland vegetation, and a larger temperature range can improve the water use efficiency of vegetation; isotherm can affect the diversity of grassland vegetation; the seasonal temperature variation coefficient plays an important role in the phenology of grassland vegetation, and the average temperature of the driest season is an important factor affecting vegetation survival; precipitation also plays an important role in the distribution and growth of grassland vegetation. For example, annual average precipitation limits the distribution range of grassland vegetation, and precipitation during the growing season determines whether grassland locusts have sufficient food sources. The precipitation variation coefficient and total precipitation affect the productivity of plant communities; topographic factors also play an important role in the distribution of grassland vegetation. Altitude indirectly affects vegetation productivity, distribution characteristics, and biomass by influencing temperature changes, while slope interacts with soil to jointly affect vegetation biomass; land use type is also an environmental factor affecting the distribution of grassland vegetation.

[0047] Therefore, based on the above analysis, the second monitoring indicator in the monitoring index set for suitable locust host areas includes at least one of the following: annual average temperature, daily average temperature range, isotherm, temperature seasonality coefficient, annual average precipitation, precipitation variation coefficient, total precipitation, precipitation during the locust host's growing season, altitude, slope, aspect, and land use type.

[0048] S120: Input the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set into the pre-trained suitable habitat monitoring model to obtain the locust suitable habitat monitoring results of the target area in the future within a preset time period.

[0049] Among them, the suitable habitat monitoring model is trained based on the sample locust suitable habitat monitoring index set and the sample locust host suitable habitat monitoring index set of the sample area.

[0050] The locust suitable habitat monitoring method provided in this embodiment no longer relies solely on meteorological data as a monitoring indicator. Instead, it fully considers multiple environmental factors affecting locust growth and development. A set of monitoring indicators for locust suitable habitats in the target area is constructed based on these factors. Furthermore, considering that changes in locust suitable habitats are also influenced by the distribution of locust hosts, a set of monitoring indicators for locust host suitable habitats in the target area is also constructed based on these factors. These two sets of indicators are then input into a pre-trained suitable habitat monitoring model to obtain the locust suitable habitat monitoring results for the target area within a preset future time period. This avoids monitoring errors caused by a single indicator, effectively improves the accuracy of locust suitable habitat monitoring results, and enables precise dynamic monitoring of future locust suitable habitats.

[0051] In some embodiments, the suitable habitat monitoring model is a heterogeneous ensemble model constructed based on the RF algorithm, XGBoost algorithm, and LightGBM algorithm. The locust suitable habitat monitoring results are generated by the suitable habitat monitoring model based on the following steps: First suitable habitat monitoring results are generated by using the RF algorithm to monitor suitable habitats based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set; second suitable habitat monitoring results are generated by using the XGBoost algorithm to monitor suitable habitats based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set; third suitable habitat monitoring results are generated by using the LightGBM algorithm to monitor suitable habitats based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set; and the average of the first, second, and third suitable habitat monitoring results is calculated to generate the locust suitable habitat monitoring results.

[0052] Machine learning models possess powerful nonlinear modeling capabilities and high adaptability, enabling them to extract latent features from high-dimensional and complex data. Therefore, applying machine learning models to the monitoring and early warning of locust habitats can effectively improve monitoring efficiency. While using a single machine learning model for habitat monitoring is easy to implement, it also has significant limitations: First, single models are often constrained by the linear separability assumption of linear models, making it difficult to fully capture the nonlinear relationships among the multiple environmental factors affecting the growth, development, and reproduction of grassland locusts and their hosts. Second, single models are prone to underfitting and overfitting due to bias and variance; simple models may underfit due to high bias, while complex models may overfit due to high variance. Therefore, improvements to existing machine learning models are necessary.

[0053] This embodiment addresses the need for dynamic monitoring of suitable habitats for grassland locusts and their host plants under future climate scenarios. It comprehensively considers factors such as the model's long-term suitable habitat monitoring capabilities, the high-dimensionality of multi-source heterogeneous data, the data volume of future climate data and remote sensing data, and computation time. A heterogeneous ensemble model framework integrating the RF (Random Forest) algorithm, the XGBoost (eXtreme Gradient Boosting) algorithm, and the LightGBM (Light Gradient Boosting Machine) algorithm is constructed. In other words, the suitable habitat monitoring model is a heterogeneous ensemble model built based on the RF algorithm, the XGBoost algorithm, and the LightGBM algorithm.

[0054] Specifically, after inputting the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set into the suitable habitat monitoring model, the model can use the RF algorithm to monitor the suitable habitat based on these two sets, generating the first suitable habitat monitoring result. .

[0055] The RF algorithm can perform sampling with replacement on data such as the distribution points of grassland locusts and the distribution points of locust hosts within the target area, as well as each monitoring indicator in the locust suitable habitat monitoring indicator set and the locust host suitable habitat monitoring indicator set. Each sampling forms a decision tree, and finally, the results of the multiple decision trees are voted on to generate the first suitable habitat monitoring result. This approach not only reduces the risk of overfitting but also reduces random errors, exhibiting good robustness.

[0056] Furthermore, the suitable habitat monitoring model can use the XGBoost algorithm to monitor suitable habitats based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set, generating a second suitable habitat monitoring result. .

[0057] The XGBoost algorithm employs a boosting framework to progressively train decision trees and continuously correct errors in the previous decision tree, ultimately resulting in an accurate decision tree. This method significantly improves the model's ability to monitor and warn of suitable habitats and enables the model to process large amounts of data.

[0058] Furthermore, the suitable habitat monitoring model can use the LightGBM algorithm to monitor suitable habitats based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set, generating a third suitable habitat monitoring result. .

[0059] The LightGBM algorithm also employs the Boosting framework and can discretize the environmental factors affecting grassland locusts and their hosts based on histogram algorithms. This method can improve the computational efficiency of the model.

[0060] Furthermore, based on the monitoring results of the first, second, and third suitable habitat areas, an average calculation is performed to generate the monitoring results of locust suitable habitat areas. .

[0061] Among them, monitoring results of locust suitable habitat areas The expression is as follows: .

[0062] The locust habitat monitoring method provided in this embodiment fully utilizes the high accuracy and robustness of the RF algorithm in habitat monitoring, as well as the powerful processing capabilities and high efficiency of the XGBoost and LightGBM algorithms for large amounts of remote sensing and meteorological data. By integrating these three algorithms, a heterogeneous ensemble model capable of monitoring locust habitats is constructed. This model generates the final locust habitat monitoring results by fusing the monitoring results from the three heterogeneous algorithms. This effectively compensates for the shortcomings of using a single model, integrates the complementary advantages of different algorithms, and enhances the model's ability to analyze complex multi-source heterogeneous influencing factors and comprehensively model complex patterns. Simultaneously, the model can improve its generalization ability through a synergistic mechanism that balances bias and variance, which is beneficial for improving the accuracy of locust habitat monitoring results.

[0063] In some embodiments, the locust suitable habitat monitoring index set includes at least one first monitoring index that affects locust growth and development; the locust suitable habitat monitoring index set is determined based on the following steps: obtaining multiple initial first monitoring indicators; performing a difference test on the multiple initial first monitoring indicators to obtain a first difference test result, and selecting at least two candidate first monitoring indicators from the multiple initial first monitoring indicators based on the first difference test result; the candidate first monitoring indicators are monitoring indicators whose influence on locust growth and development is greater than or equal to a first preset threshold; performing a correlation analysis on all candidate first monitoring indicators to obtain a first correlation analysis result, and selecting at least one first monitoring indicator from all candidate first monitoring indicators based on the first correlation analysis result and adding it to the locust suitable habitat monitoring index set.

[0064] Understandably, there are various monitoring indicators that affect locust growth and development, but not all monitoring indicators that affect locust growth and development are applicable to the model's monitoring of changes in locust habitat. Therefore, monitoring indicators that do not have a significant impact on changes in locust habitat need to be excluded.

[0065] Specifically, the first step is to obtain all initial monitoring indicators that affect the growth and development of locusts.

[0066] Furthermore, the Mann-Whitney U test method was used to conduct a difference test on all initial first monitoring indicators to obtain the first difference test results. Based on the first difference test results, monitoring indicators that did not have a significant impact on the changes in locust habitat were excluded from all initial first monitoring indicators, and at least two candidate first monitoring indicators were retained. The candidate first monitoring indicators are actually monitoring indicators whose impact on locust growth and development is greater than or equal to the first preset threshold.

[0067] Optionally, such as Figure 2 As shown, the initial first monitoring indicator affecting locust growth and development may be a dynamic variable indicator or a static variable indicator. Depending on the type of indicator, the initial first monitoring indicator can be tested for differences in dynamic or static variables.

[0068] For example, the seasonal variation coefficient of temperature (TS), the coefficient of variation of precipitation (PS), the precipitation during the egg stage (EP), and the precipitation during the locust nymph stage (NP) of the target area are all dynamic variables, and their differences can be tested. On the other hand, the elevation and slope of the target area are static variables, and their differences can be tested.

[0069] Furthermore, such as Figure 3 As shown, a correlation analysis was performed on all candidate first monitoring indicators to obtain the first correlation analysis results. Based on the first correlation analysis results, at least one first monitoring indicator was selected from all candidate first monitoring indicators and added to the locust suitable habitat monitoring indicator set.

[0070] The first correlation analysis results include the correlation between any two candidate first monitoring indicators.

[0071] Generally, if there is a strong correlation between two monitoring indicators, that is, if there is a strong correlation between the two monitoring indicators... If the value exceeds a preset threshold (e.g., 0.8), it can lead to collinearity issues in the model, reducing its generalization ability and monitoring performance. Therefore, it is necessary to exclude some monitoring indicators with strong correlations.

[0072] Specifically, for any two candidate first monitoring indicators, if the correlation between these two candidate first monitoring indicators... If the score is greater than the preset threshold, the two candidate first monitoring indicators can be tested for importance using the random forest Gini importance test method. The feature importance scores corresponding to the two candidate first monitoring indicators are calculated respectively. Based on the feature importance scores corresponding to the two candidate first monitoring indicators, the indicators are screened, and only the candidate first monitoring indicators with high feature importance scores are retained as the first monitoring indicators.

[0073] In the process of feature importance assessment, n (n is a positive integer) samples can be randomly selected from the occurrence and non-occurrence points of grassland locusts and their hosts in the target area to construct a decision tree. The decision tree includes multiple nodes. Then, each candidate first monitoring indicator is regarded as a feature. For each feature, the importance of the feature is assessed by calculating the reduction in Gini impurity when the feature splits at the decision tree node.

[0074] Here, a feature corresponds to a node in the decision tree. For a node S in the decision tree, its Gini impurity is defined as: ; in, Indicates the Gini impurity of node S; Indicates the first node The proportion of class samples; This indicates the total number of sample categories.

[0075] During the splitting process of the decision tree, the first... Features When performing node splitting, the reduction in Gini impurity before and after the split can be calculated. For the parent node... In other words, if the parent node Split into two child nodes and Then the parent node The reduction in impurity of the ginni for: ; in, Indicates the parent node The impurity of the gin; Indicates the left child node The impurity of the gin; Indicates the right child node The impurity of the gin; The weight representing the number of samples in the left child node; The weight represents the number of samples in the right child node.

[0076] Construct multiple decision trees, targeting the first... Features The final Gini importance is the average reduction in Gini impurity of this feature across all decision trees. , The expression is as follows: ; in, Indicates the total number of decision trees; For the first Features in a decision tree The amount of Gini reduction during each split; Indicates the number of splits.

[0077] For each feature, after calculating the final Gini importance of that feature, the final Gini importance is used as the feature importance score of the corresponding candidate first monitoring indicator.

[0078] For any two candidate first monitoring indicators that are strongly correlated, the feature importance scores of these two candidate first monitoring indicators can be compared. The candidate first monitoring indicators with low feature importance scores are excluded, and only the candidate first monitoring indicators with high feature importance scores are retained as the first monitoring indicators. The first monitoring indicator is then added to the monitoring indicator set for locust suitable habitat areas.

[0079] Optionally, such as Figure 4 As shown, the elevation, slope, seasonal temperature variation coefficient (TS), precipitation variation coefficient (PS), egg stage precipitation (EP), and locust nymph stage precipitation (NP) of the target area can be used as the primary monitoring indicators.

[0080] In some embodiments, the first monitoring indicator in the locust suitable habitat monitoring indicator set includes at least one of the following in the target area: egg stage precipitation, locust nymph stage precipitation, temperature, altitude, slope, isotherm, temperature seasonal variation coefficient, precipitation variation coefficient, pre-hatching precipitation of locusts, temperature and precipitation during the locust oviposition period, and land use type.

[0081] In some embodiments, the locust host suitable habitat monitoring index set includes at least one second monitoring index that affects the distribution of locust hosts; the locust host suitable habitat monitoring index set is determined based on the following steps: obtaining multiple initial second monitoring indicators; performing a difference test on the multiple initial second monitoring indicators to obtain a second difference test result, and selecting at least two candidate second monitoring indicators from the multiple initial second monitoring indicators based on the second difference test result; the candidate second monitoring indicators are monitoring indicators whose influence on the distribution of locust hosts is greater than or equal to a second preset threshold; performing a correlation analysis on all candidate second monitoring indicators to obtain a second correlation analysis result, and selecting at least one second monitoring indicator from all candidate second monitoring indicators based on the second correlation analysis result and adding it to the locust host suitable habitat monitoring index set.

[0082] Similarly, there are various monitoring indicators that affect the distribution of locust hosts, but not all of these indicators have statistically significant effects on vegetation distribution. Therefore, monitoring indicators that do not have a significant effect need to be excluded.

[0083] Specifically, we first obtain all initial secondary monitoring indicators that affect the distribution of locust hosts.

[0084] Furthermore, the Mann-Whitney U test method was used to test the difference of all initial second monitoring indicators to obtain the second difference test results. Based on the second difference test results, monitoring indicators with no significant difference were excluded from all initial second monitoring indicators, and at least two candidate second monitoring indicators were retained. The candidate second monitoring indicators are monitoring indicators whose influence on the distribution of locust hosts is greater than or equal to the second preset threshold.

[0085] Optionally, such as Figure 5 As shown, the initial second monitoring indicator affecting the distribution of locust hosts may be a dynamic variable indicator or a static variable indicator. Depending on the type of indicator, the initial second monitoring indicator can be tested for dynamic variable differences or static variable differences.

[0086] Furthermore, such as Figure 6 As shown, the Spearman method was used to perform correlation analysis on all candidate second monitoring indicators to obtain the second correlation analysis results. Based on the second correlation analysis results, at least one second monitoring indicator was selected from all candidate second monitoring indicators and added to the monitoring indicator set for suitable locust host areas.

[0087] The second correlation analysis results include the correlation between any two candidate second monitoring indicators.

[0088] Generally, if there is a strong correlation between two monitoring indicators, that is, if there is a strong correlation between the two monitoring indicators... If the value exceeds a preset threshold (e.g., 0.8), it can lead to collinearity issues in the model, reducing its generalization ability and monitoring performance. Therefore, it is necessary to exclude some monitoring indicators with strong correlations.

[0089] Specifically, for any two candidate second monitoring indicators, if the correlation between these two candidate second monitoring indicators... If the score is greater than the preset threshold, the two candidate second monitoring indicators can be tested for importance using the random forest Gini importance test method. The feature importance scores corresponding to the two candidate second monitoring indicators are calculated respectively. Based on the feature importance scores corresponding to the two candidate second monitoring indicators, the indicators are screened, and only the candidate second monitoring indicators with high feature importance scores are retained as second monitoring indicators and added to the locust host suitable habitat monitoring indicator set.

[0090] Optionally, such as Figure 7 As shown, the target area's elevation, slope, seasonal temperature variation (TS), host vegetation growing season precipitation (GP), and isotherm (IST) can be used as secondary monitoring indicators.

[0091] In some embodiments, the second monitoring indicator in the locust host suitable habitat monitoring indicator set includes at least one of the following for the target area: annual average temperature, daily average temperature range, isotherm, temperature seasonality coefficient, annual average precipitation, precipitation variation coefficient, total precipitation, locust host growth period precipitation, altitude, slope, aspect, and land use type.

[0092] Compared with existing technologies, the locust habitat monitoring method provided by this invention has at least the following advantages: (1) Stronger element synergy: This invention breaks through the limitation of using only a single meteorological factor in traditional methods, and comprehensively considers the synergistic effect of multiple source elements such as meteorology, vegetation, land use type and topography, so that the model can learn and understand the complex environmental mechanism of locust plague more comprehensively; (2) Clear mechanism orientation: Based on the key environmental factors of different growth and development stages of locusts, a locust habitat monitoring index system is constructed, so that the model has clear ecological mechanism support, which is conducive to improving the reliability and interpretability of the model monitoring results; (3) Realize the coupling analysis of host and locust: Couple the dynamics of host habitat in future scenarios, so that the model can understand the key constraints of parasitic plants on locust distribution and improve the model's understanding of biological interactions.

[0093] The locust habitat monitoring method provided by this invention has been verified in training data under various different climatic scenarios such as SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5. Experiments show that the locust habitat monitoring method provided by this invention can achieve scientific and accurate dynamic monitoring of habitats.

[0094] This invention also provides a device for monitoring locust habitats. Please refer to [link / reference]. Figure 8 , Figure 8 This is a schematic diagram of the locust habitat monitoring device provided by the present invention. In this embodiment, the locust habitat monitoring device includes an acquisition module 810 and a habitat monitoring module 820.

[0095] The acquisition module 810 is used to acquire the monitoring index set of locust suitable habitat area and the monitoring index set of locust host suitable habitat area in the target area.

[0096] The monitoring index set for suitable locust habitats is determined based on environmental factors that affect locust growth and development, while the monitoring index set for suitable locust host habitats is determined based on environmental factors that affect the distribution of locust hosts. These environmental factors include climate factors, vegetation factors, land use type factors, and topographic factors.

[0097] The suitable habitat monitoring module 820 is used to input the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set into the pre-trained suitable habitat monitoring model to obtain the locust suitable habitat monitoring results of the target area in the future within a preset time period.

[0098] Among them, the suitable habitat monitoring model is trained based on the sample locust suitable habitat monitoring index set and the sample locust host suitable habitat monitoring index set of the sample area.

[0099] In some embodiments, the suitable habitat monitoring model is a heterogeneous ensemble model constructed based on the RF algorithm, XGBoost algorithm, and LightGBM algorithm. The locust suitable habitat monitoring results are generated by the suitable habitat monitoring model based on the following steps: First suitable habitat monitoring results are generated by using the RF algorithm to monitor suitable habitats based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set; second suitable habitat monitoring results are generated by using the XGBoost algorithm to monitor suitable habitats based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set; third suitable habitat monitoring results are generated by using the LightGBM algorithm to monitor suitable habitats based on the locust suitable habitat monitoring index set and the locust host suitable habitat monitoring index set; and the average of the first, second, and third suitable habitat monitoring results is calculated to generate the locust suitable habitat monitoring results.

[0100] In some embodiments, the locust suitable habitat monitoring index set includes at least one first monitoring index that affects locust growth and development; the locust suitable habitat monitoring index set is determined based on the following steps: obtaining multiple initial first monitoring indicators; performing a difference test on the multiple initial first monitoring indicators to obtain a first difference test result, and selecting at least two candidate first monitoring indicators from the multiple initial first monitoring indicators based on the first difference test result; the candidate first monitoring indicators are monitoring indicators whose influence on locust growth and development is greater than or equal to a first preset threshold; performing a correlation analysis on all candidate first monitoring indicators to obtain a first correlation analysis result, and selecting at least one first monitoring indicator from all candidate first monitoring indicators based on the first correlation analysis result and adding it to the locust suitable habitat monitoring index set.

[0101] In some embodiments, the first monitoring indicator in the locust habitat monitoring indicator set includes at least one of the following in the target area: egg stage precipitation, locust nymph stage precipitation, temperature, altitude, slope, isotherm, temperature seasonal variation coefficient, precipitation variation coefficient, pre-hatching precipitation of locusts, temperature and precipitation during the locust oviposition period, and land use type.

[0102] In some embodiments, the locust host suitable habitat monitoring index set includes at least one second monitoring index that affects the distribution of locust hosts; the locust host suitable habitat monitoring index set is determined based on the following steps: obtaining multiple initial second monitoring indicators; performing a difference test on the multiple initial second monitoring indicators to obtain a second difference test result, and selecting at least two candidate second monitoring indicators from the multiple initial second monitoring indicators based on the second difference test result; the candidate second monitoring indicators are monitoring indicators whose influence on the distribution of locust hosts is greater than or equal to a second preset threshold; performing a correlation analysis on all candidate second monitoring indicators to obtain a second correlation analysis result, and selecting at least one second monitoring indicator from all candidate second monitoring indicators based on the second correlation analysis result and adding it to the locust host suitable habitat monitoring index set.

[0103] In some embodiments, the second monitoring indicator in the locust host suitable habitat monitoring indicator set includes at least one of the following for the target area: annual average temperature, daily average temperature range, isotherm, temperature seasonality coefficient, annual average precipitation, precipitation variation coefficient, total precipitation, locust host growth period precipitation, altitude, slope, aspect, and land use type.

[0104] The present invention also provides an electronic device. Figure 9 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 9As shown, the electronic device may include a processor 910, a communication interface 920, a memory 930, and a communication bus 940. The processor 910, communication interface 920, and memory 930 communicate with each other via the communication bus 940. The processor 910 can call logical instructions from the memory 930 to execute a locust habitat monitoring method.

[0105] Furthermore, the logical instructions in the aforementioned memory 930 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0106] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the locust habitat monitoring method provided by the above methods.

[0107] The present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can perform the locust habitat monitoring method provided by the above methods.

[0108] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0109] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method of monitoring a locust habitat, characterized by, The method comprises: obtaining a locust suitable habitat monitoring indicator set of a target area and a locust host suitable habitat monitoring indicator set; the locust suitable habitat monitoring indicator set is determined based on environmental factors affecting the growth and development of locusts, and the locust host suitable habitat monitoring indicator set is determined based on environmental factors affecting the distribution of locust hosts, and the environmental factors include climate factors, vegetation factors, land use type factors and terrain factors; inputting the locust suitable habitat monitoring indicator set and the locust host suitable habitat monitoring indicator set into a pre-trained suitable habitat monitoring model to obtain a locust suitable habitat monitoring result of the target area in a future preset time period output by the suitable habitat monitoring model; wherein the suitable habitat monitoring model is trained based on a sample locust suitable habitat monitoring indicator set and a sample locust host suitable habitat monitoring indicator set of a sample area.

2. The method of claim 1, wherein, The suitable habitat monitoring model is a heterogeneous ensemble model constructed based on RF algorithm, XGBoost algorithm and LightGBM algorithm; The locust suitable habitat monitoring result is generated by the suitable habitat monitoring model based on the following steps: based on the locust suitable habitat monitoring indicator set and the locust host suitable habitat monitoring indicator set, performing suitable habitat monitoring through the RF algorithm to generate a first suitable habitat monitoring result; based on the locust suitable habitat monitoring indicator set and the locust host suitable habitat monitoring indicator set, performing suitable habitat monitoring through the XGBoost algorithm to generate a second suitable habitat monitoring result; based on the locust suitable habitat monitoring indicator set and the locust host suitable habitat monitoring indicator set, performing suitable habitat monitoring through the LightGBM algorithm to generate a third suitable habitat monitoring result; based on the first suitable habitat monitoring result, the second suitable habitat monitoring result and the third suitable habitat monitoring result, performing mean operation to generate the locust suitable habitat monitoring result.

3. The method of claim 1, wherein, The locust suitable habitat monitoring indicator set includes at least one first monitoring indicator affecting the growth and development of locusts; The locust suitable habitat monitoring indicator set is determined based on the following steps: obtaining a plurality of initial first monitoring indicators; performing difference test on a plurality of the initial first monitoring indicators to obtain a first difference test result, and based on the first difference test result, selecting at least two candidate first monitoring indicators from a plurality of the initial first monitoring indicators; the candidate first monitoring indicator is a monitoring indicator whose influence on the growth and development of locusts is greater than or equal to a first preset threshold; performing correlation analysis on all the candidate first monitoring indicators to obtain a first correlation analysis result, and based on the first correlation analysis result, selecting at least one first monitoring indicator from all the candidate first monitoring indicators and adding it to the locust suitable habitat monitoring indicator set.

4. The method of claim 3, wherein, The first monitoring indicator in the locust suitable habitat monitoring indicator set includes at least one of the following: precipitation in the egg stage of the target area, precipitation in the nymph stage, temperature, altitude, slope, isothermality, temperature seasonal variation coefficient, precipitation variation coefficient, pre-hatching precipitation of locusts, temperature and precipitation in the locust oviposition period, and land use type.

5. The method of claim 1, wherein, The second monitoring index set of the locust host suitable area includes at least one second monitoring index affecting the distribution of the locust host; The second monitoring index set of the locust host suitable area is determined based on the following steps: Obtaining a plurality of initial second monitoring indexes; Performing difference test on the plurality of initial second monitoring indexes to obtain a second difference test result, and selecting at least two candidate second monitoring indexes from the plurality of initial second monitoring indexes based on the second difference test result; The candidate second monitoring index is a monitoring index whose influence degree on the distribution of the locust host is greater than or equal to a second preset threshold; Performing correlation analysis on all the candidate second monitoring indexes to obtain a second correlation analysis result, and selecting at least one second monitoring index from all the candidate second monitoring indexes based on the second correlation analysis result and adding the second monitoring index to the second monitoring index set of the locust host suitable area.

6. The method of claim 5, wherein, The second monitoring index in the second monitoring index set of the locust host suitable area includes at least one of the annual average temperature, the average daily temperature range, the isothermality, the temperature seasonal variation coefficient, the annual average precipitation, the precipitation variation coefficient, the total precipitation, the precipitation during the growth period of the locust host, the altitude, the slope, the slope direction, and the land use type of the target area.

7. A locust habitat monitoring device, characterized by, Comprise: An acquisition module is configured to acquire a locust suitable area monitoring index set and a locust host suitable area monitoring index set of a target area; the locust suitable area monitoring index set is determined based on environmental factors affecting the growth and development of locusts, and the locust host suitable area monitoring index set is determined based on environmental factors affecting the distribution of locust hosts; the environmental factors include climate factors, vegetation factors, land use type factors, and topographic factors; A suitable area monitoring module is configured to input the locust suitable area monitoring index set and the locust host suitable area monitoring index set into a pre-trained suitable area monitoring model to obtain a locust suitable area monitoring result of the target area in a future preset time period output by the suitable area monitoring model; The suitable area monitoring model is trained based on a sample locust suitable area monitoring index set and a sample locust host suitable area monitoring index set of a sample area.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and running on the processor, characterized in that, The processor executes the computer program to implement the locust suitable area monitoring method of any one of claims 1 to 6. 9.A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the locust suitable area monitoring method of any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the locust suitable area monitoring method of any one of claims 1 to 6.

Images