Fine economic forest fruit flowering phase low-temperature cold injury risk assessment and early warning method and device

Abstract

The invention provides a refined economic forest fruit flowering phase low-temperature cold damage risk assessment early warning method and device, and relates to the technical field of disaster assessment, and the method comprises the steps: obtaining a multi-source historical data set corresponding to a research region, and planting economic forest fruits in the research region; preprocessing the multi-source historical data set to obtain a characteristic quantity matched with a forest fruit growth area and a forest fruit flowering phase of the economic forest fruit, and determining an event level corresponding to a low-temperature cold injury event occurring in the research area based on preprocessed historical single-station low-temperature cold injury event data; and based on the characteristic quantity of the economic forest fruit and the event level corresponding to the low-temperature cold injury event, training the neural network model to obtain a cold injury risk assessment and early warning model, the cold injury risk assessment and early warning model being used for performing assessment and early warning on the event level corresponding to the low-temperature cold injury event occurring in the forest fruit growth area in the forest fruit flowering phase. According to the invention, the precision degree and accuracy of low-temperature cold injury risk assessment early warning can be obviously improved.

Description

Technical Field

[0001] This invention relates to the field of disaster assessment technology, and in particular to a refined method and device for assessing and warning of low-temperature chilling damage during the flowering period of economic forest fruits. Background Technology

[0002] Low-temperature chilling injury is a common natural disaster in economic forest and fruit production, significantly impacting flowering, fruit development, and yield, and potentially leading to severe economic losses. Therefore, conducting research on risk assessment and early warning of low-temperature chilling injury during the flowering period of economic forest and fruit trees is crucial for guiding production practices and mitigating disaster losses. Traditional methods for risk assessment and early warning of low-temperature chilling injury in economic forest and fruit trees mainly include statistical analysis based on historical data, predictive simulation based on physical models, and expert assessment based on experience. While these methods can provide valuable references for risk assessment and early warning of low-temperature chilling injury to some extent, they either rely heavily on historical data, making it difficult to fully capture the nonlinear relationships within the data, or they involve complex parameters, high computational demands, and are easily affected by various uncertainties.

[0003] Therefore, in the current era of rapid development of artificial intelligence algorithms, AI methods possess stronger generalization capabilities, better adaptability, and more comprehensive considerations. Currently, AI-based methods for assessing and warning of low-temperature damage risks primarily rely on ground-based observation data such as temperature and humidity collected from meteorological stations. These data are then used to generate regional risk distribution maps through spatial interpolation algorithms. However, due to the limited number and uneven spatial distribution of ground-based meteorological stations, especially in complex terrain areas, relying solely on station interpolation methods is insufficient to accurately reflect small-scale climate characteristics, leading to significant errors in risk assessment results. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a refined method and device for assessing and warning of low-temperature chilling injury risk during the flowering period of economic forest fruits, which can significantly improve the precision and accuracy of low-temperature chilling injury risk assessment and warning.

[0005] In a first aspect, the present invention provides a refined method for assessing and warning of low-temperature chilling injury risk during the flowering period of economic forest fruit trees, including: The study area was equipped with a multi-source historical dataset, which included historical single-station low-temperature cold damage event data, ground meteorological station observation data, geographic feature data, and satellite observation data. The multi-source historical dataset was preprocessed to obtain characteristic quantities that match the fruit growth area and flowering period of economic forests and fruits. Based on the preprocessed historical single-station low-temperature chilling injury event data, the event level corresponding to the low-temperature chilling injury events that occurred in the study area was determined. Based on the characteristic quantities of economic forest fruits and the event levels corresponding to low temperature chilling injury events, a neural network model is trained to obtain a chilling injury risk assessment and early warning model. The chilling injury risk assessment and early warning model is used to assess and warn of the event levels corresponding to low temperature chilling injury events occurring in forest fruit growing areas during the flowering period.

[0006] In one implementation, the multi-source historical data set is preprocessed to obtain feature quantities that match the fruit tree growth area and flowering period of economic fruit trees, including: Based on the flowering period of economic forest fruits, time matching was performed on historical single-station low-temperature cold damage event data and ground meteorological station observation data in the multi-source historical dataset; based on the forest fruit growth area of ​​economic forest fruits, spatial matching was performed on geographical feature data and satellite observation data in the multi-source historical dataset. Preprocessing was performed on the matched historical single-station low-temperature chilling injury event data, ground meteorological station observation data, geographic feature data, and satellite observation data to obtain characteristic quantities that match the fruit growth area and flowering period of economic forests and fruits. The characteristic quantities include the statistics of historical single-station low-temperature chilling injury event data and meteorological station observation data, geographic feature data, and surface temperature data.

[0007] In one implementation, based on preprocessed historical single-station low-temperature chilling injury event data, the event level corresponding to the low-temperature chilling injury events occurring within the study area is determined, including: Based on the preprocessed historical single-station low-temperature chilling injury event data, the characteristic quantities of low-temperature chilling injury events during each fruit and forest flowering period are statistically analyzed. The characteristic quantities of low-temperature chilling injury events include one or more of the following: number of days affected by chilling injury, minimum temperature during chilling injury, average temperature during chilling injury, maximum chilling injury intensity, and average chilling injury intensity. Based on the pre-set cold damage risk index system and the characteristic quantities of low-temperature cold damage events, the event level corresponding to the low-temperature cold damage events occurring in the study area is determined.

[0008] In one implementation, the neural network employs a multilayer perceptron model.

[0009] In one implementation, a neural network model is trained based on the characteristic quantities of economic forest fruits and the event levels corresponding to low-temperature chilling injury events, including: A sample set is constructed based on the characteristics of economic forest fruits. The event level corresponding to the low temperature damage event is used as the sample label. The sample set and its corresponding sample label are divided into a training set and a test set. The training set is used to train the multilayer perceptron model.

[0010] In one implementation, the method further includes: Using the test set, the mean absolute error and root mean square error of the cold damage risk assessment and early warning model are determined to evaluate the quality of the model.

[0011] In one implementation, the method further includes: Obtain satellite or drone observation results corresponding to the study area to determine whether the economic fruit trees planted in the study area are in the flowering period; If so, based on the fruit-growing area of ​​economic forests, extract the corresponding feature quantities of economic forests from the multi-source current dataset corresponding to the study area; The chilling injury risk assessment and early warning model is used to assess and warn of the event level of low temperature chilling injury events occurring in the fruit growing area based on the characteristic quantities corresponding to economic forests and fruits.

[0012] Secondly, the present invention also provides a refined low-temperature chilling injury risk assessment and early warning device for economic forest fruit trees during flowering, comprising: The data acquisition module is used to acquire multi-source historical data sets corresponding to the study area. The study area is planted with economic forests and fruit trees. The multi-source historical data sets include historical single-station low temperature cold damage event data, ground meteorological station observation data, geographic feature data and satellite observation data. The data processing module is used to preprocess the multi-source historical data set to obtain feature quantities that match the fruit growth area and flowering period of economic forests and fruits, and to determine the event level corresponding to the low temperature damage events that occurred in the study area based on the preprocessed historical single-station low temperature damage event data. The model training module is used to train the neural network model based on the characteristic quantities of economic forest fruits and the event level corresponding to low temperature chilling injury events, so as to obtain the chilling injury risk assessment and early warning model. The chilling injury risk assessment and early warning model is used to assess and warn of the event level corresponding to low temperature chilling injury events occurring in forest fruit growing areas during the flowering period.

[0013] Thirdly, the present invention also provides an electronic device including a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement any of the methods provided in the first aspect.

[0014] Fourthly, the present invention also provides a computer-readable storage medium storing computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement any of the methods provided in the first aspect.

[0015] This invention provides a refined method and device for assessing and warning of low-temperature chilling injury risk during the flowering period of economic fruit trees. First, a multi-source historical data set corresponding to the study area is acquired. The study area contains economic fruit trees, and the multi-source historical data set includes historical single-station low-temperature chilling injury event data, ground meteorological station observation data, geographical feature data, and satellite observation data. Then, the multi-source historical data set is preprocessed to obtain feature quantities matching the fruit tree growth area and flowering period of the economic fruit trees. Based on the preprocessed historical single-station low-temperature chilling injury event data, the event level corresponding to the low-temperature chilling injury events occurring in the study area is determined. Finally, based on the feature quantities of the economic fruit trees and the event levels corresponding to the low-temperature chilling injury events, a neural network model is trained to obtain a chilling injury risk assessment and warning model. This model is used to assess and warn of the event level corresponding to low-temperature chilling injury events occurring in the fruit tree growth area during the flowering period. The above method utilizes multi-source data, including historical single-station low-temperature chilling injury event data, ground meteorological station observation data, geographical feature data, and satellite observation data, combined with the growth area and flowering period of economic forest fruits. It leverages the powerful learning and predictive capabilities of neural network models to conduct early assessment and warning of the risks of low-temperature chilling injury to economic forest fruits during their flowering period.

[0016] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described in detail below with reference to the accompanying drawings. Attached Figure Description

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

[0019] Figure 1 A flowchart illustrating a refined method for assessing and warning of low-temperature chilling injury during the flowering period of economic fruit trees, provided in an embodiment of the present invention. Figure 2 This is a technical framework diagram of a refined method for assessing and warning of low-temperature chilling injury during the flowering period of economic forest fruit trees, provided by an embodiment of the present invention. Figure 3 This is a schematic diagram of the structure of a refined low-temperature chilling injury risk assessment and early warning device for fruit trees during flowering period provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Currently, existing risk assessment methods suffer from significant errors in their results. To address this, this invention provides a refined method and device for assessing and warning of low-temperature chilling injury risks during the flowering period of economic forest fruits. This significantly improves the precision and accuracy of low-temperature chilling injury risk assessment and warning. Specifically, based on artificial intelligence methods combined with advanced satellite remote sensing technology and gridded temperature prediction data, a low-temperature chilling injury risk assessment and warning model for economic forest fruits is constructed. This model provides a refined risk map for the fruit flowering period 1-3 days in advance. It reduces reliance on cumbersome data, comprehensively captures implicit features within the data, and lowers the uncertainty of risk prediction. Simultaneously, it enhances the application value of artificial intelligence technology in production practice, providing farmers, economic forest fruit planting enterprises, agricultural disaster insurance, and related departments with more accurate and reliable forecasting and decision-making support.

[0022] To facilitate understanding of this embodiment, a detailed description of the refined low-temperature chilling injury risk assessment and early warning method for economic forest fruit trees during flowering period, as disclosed in this embodiment of the invention, will be provided first. (See [link to relevant documentation]). Figure 1 The diagram shows a flowchart of a refined method for assessing and warning of low-temperature chilling injury during the flowering period of fruit trees in economic forests. The method mainly includes the following steps S102 to S106: Step S102: Obtain the multi-source historical data set corresponding to the study area.

[0023] The study area includes economic fruit trees, and the multi-source historical data set includes historical single-station low-temperature cold damage event data, ground meteorological station observation data, geographic feature data, and satellite observation data. Step S104: Preprocess the multi-source historical data set to obtain feature quantities that match the fruit growth area and flowering period of economic forests and fruits, and determine the event level corresponding to the low temperature damage events that occurred in the study area based on the preprocessed historical single-station low temperature damage event data.

[0024] Among them, the characteristic quantities of economic forest fruits include historical single-station low temperature cold damage event data, statistical quantities of meteorological station observation data, geographical feature data, and surface temperature data.

[0025] In one implementation, the preprocessing procedure is as follows: Based on the flowering period of economic forest fruits, time matching was performed on historical single-station low-temperature cold damage event data and ground meteorological station observation data in the multi-source historical dataset; based on the forest fruit growth area of ​​economic forest fruits, spatial matching was performed on geographical feature data and satellite observation data in the multi-source historical dataset. Preprocessing was performed on the matched historical single-station low-temperature chilling injury event data, ground meteorological station observation data, geographic feature data, and satellite observation data to obtain characteristic quantities that match the fruit growth area and flowering period of economic forests and fruits. The characteristic quantities include the statistics of historical single-station low-temperature chilling injury event data and meteorological station observation data, geographic feature data, and surface temperature data.

[0026] In one implementation, the process for determining the event level corresponding to a low-temperature chilling injury event is as follows: Based on the preprocessed historical single-station low-temperature chilling injury event data, the characteristic quantities of low-temperature chilling injury events during each fruit and forest flowering period are statistically analyzed. The characteristic quantities of low-temperature chilling injury events include one or more of the following: number of days affected by chilling injury, minimum temperature during chilling injury, average temperature during chilling injury, maximum chilling injury intensity, and average chilling injury intensity. Based on the pre-set cold damage risk index system and the characteristic quantities of low-temperature cold damage events, the event level corresponding to the low-temperature cold damage events occurring in the study area is determined.

[0027] Step S106: Based on the characteristic quantities of economic forest fruits and the event level corresponding to low temperature chilling injury events, the neural network model is trained to obtain a chilling injury risk assessment and early warning model. The chilling injury risk assessment and early warning model is used to assess and warn of the event level corresponding to low temperature chilling injury events occurring in forest fruit growing areas during the flowering period.

[0028] The neural network employs a multilayer perceptron model. In one example, a sample set is constructed based on the features of economic forest fruits. The event level corresponding to the low-temperature chilling injury event is used as the sample label. The sample set and its corresponding sample labels are divided into a training set and a test set. The multilayer perceptron model is trained using the training set, and the mean absolute error and root mean square error corresponding to the chilling injury risk assessment and early warning model are determined using the test set to evaluate the quality of the chilling injury risk assessment and early warning model.

[0029] In the application phase, low-temperature chilling injury early warning is conducted by combining gridded temperature forecasts for the next 1-3 days. Specifically, satellite or UAV observations of the study area are acquired to determine whether the planted economic fruit trees in the study area are in their flowering period. If so, based on the fruit tree growth area, characteristic quantities corresponding to the economic fruit trees are extracted from the multi-source current data set corresponding to the study area. The multi-source current data set includes gridded temperature forecasts for the next 1-3 days, ground meteorological station observation data, geographical feature data, and satellite observation data. Through a chilling injury risk assessment and early warning model, based on the characteristic quantities corresponding to the economic fruit trees, the event level corresponding to low-temperature chilling injury events occurring in the fruit tree growth area is assessed and early warning is issued.

[0030] The refined low-temperature chilling injury risk assessment and early warning method for economic forest fruits provided in this invention utilizes multi-source data such as historical single-station low-temperature chilling injury event data, ground meteorological station observation data, geographical feature data, and satellite observation data, combined with the growth area and flowering period of economic forest fruits, and leverages the powerful learning and predictive capabilities of neural network models to conduct early assessment and early warning of the risk of low-temperature chilling injury to economic forest fruits during their flowering period.

[0031] To facilitate understanding, this invention provides a specific implementation method for a refined low-temperature chilling injury risk assessment and early warning method for fruit trees during flowering. The overall approach is as follows: acquire historical single-station low-temperature chilling injury event data, and perform data cleaning and statistical analysis of low-temperature chilling injury characteristics during fruit tree flowering; acquire ground meteorological station data, perform spatiotemporal matching, data quality control, and statistical analysis of meteorological characteristics during fruit tree flowering; acquire geographical feature data and perform spatiotemporal matching processing; acquire relevant information such as the planting area and underlying surface type of economic fruit trees, and perform data cleaning and organization; construct a low-temperature chilling injury risk sample set for economic fruit trees, and train the model using a multilayer perceptron; and perform training and prediction based on the model. See also... Figure 2 The diagram shown is a technical framework diagram of a refined method for assessing and warning of low-temperature chilling damage during the flowering period of fruit trees in economic forests. Specifically: Step 1: Obtain historical data on single-station low-temperature chilling injury events, and perform data cleaning and statistical analysis and classification of low-temperature chilling injury characteristics during the flowering period of forest fruits: (1) Obtain historical single-station low-temperature cold damage events recorded by meteorological observation stations in the target area, and record the original information such as the station name, longitude, latitude, start date, and end date of each event; (2) Check the original information of each event, such as the station name, longitude, latitude, event start time, and event end time. Remove any information that is missing any field or has obvious outliers to complete the cleaning of the historical single-station low-temperature cold damage event dataset. (3) Based on the flowering period of each fruit tree, statistically analyze the characteristic values ​​of low temperature damage during that period, including the number of days affected by the damage, the lowest temperature during the damage process, the average temperature during the damage process, the maximum value of the damage intensity, and the average value of the damage intensity.

[0032] (4) Based on the magnitude of the cold damage characteristic values, cold damage is classified into high, relatively high, medium, relatively low, and low risk zones according to the cold damage risk index system. The classification method can be the percentile method, the standard deviation method, or the natural breakpoint method. The classification standards for the percentile method and the standard deviation method are as follows: Table 1. Percentile Ranking Standards

[0033] Table 2 Standard Deviation Method Classification Criteria

[0034] Step 2: Obtain data from ground meteorological stations, perform spatiotemporal matching, data quality control, and statistical analysis of meteorological characteristics during the flowering period of forest fruits. (1) Obtain observation data from ground meteorological stations in the target area, including national stations, regional stations and automatic stations, and the observation elements include average temperature, maximum temperature, average surface temperature, relative humidity, minimum temperature, sea level pressure, sunshine duration and average wind speed. (2) Based on the observation time of ground meteorological stations, time matching is performed with the flowering period of forest fruits in each historical period. Quality control is performed on the ground meteorological station observation data within the screened flowering period of forest fruits, including climatological boundary value check, climate extreme value check, internal consistency check, spatial consistency check and temporal consistency check, etc. (3) Statistical analysis of meteorological characteristics during the flowering period of forest fruits, such as average temperature, minimum temperature, average surface temperature, average relative humidity, total sunshine hours, average wind speed, and maximum wind speed.

[0035] Step 3: Obtain geographic feature data and perform spatiotemporal matching processing: (1) Obtain geographical feature data such as terrain data, landform type data, soil type data, and vegetation type data, and perform data analysis; (2) Based on geographic feature data, spatial matching is performed with the fruit and forest growth area. The selected geographic feature data is preprocessed by cleaning, noise reduction, and format conversion to ensure data quality and consistency.

[0036] Step 4: Acquire refined satellite data and gridded temperature forecast data for the next 1-3 days, extract surface temperature information, and perform spatiotemporal matching: (1) Acquire refined satellite data, extract surface temperature information, and perform data analysis; (2) Obtain temperature grid point forecast data for the next 1-3 days, extract temperature forecast information, and perform data analysis; (3) Spatial matching with the fruit and forest growth area, and preprocessing operations such as cleaning, noise reduction and format conversion of the selected surface temperature data to ensure the quality and consistency of the data.

[0037] Step 5: Construct a sample set of low-temperature chilling injury risk data during the flowering period of economic forest fruit trees, and train the model using a multilayer perceptron: (1) Spatial interpolation is performed on the original data of low temperature cold damage disaster events in each fruit and forest flowering period obtained in step 1, the observation statistics of ground meteorological stations in the fruit and forest flowering period obtained in step 2 (such as average temperature, minimum temperature, average surface temperature, average relative humidity, total sunshine hours, average wind speed, and maximum wind speed), the geographic information data in step 3 (topography data, landform type data, soil type data, and vegetation type data), and the surface temperature data in step 4. The above data are combined to form a sample set, in which the low temperature cold damage risk level is the sample label, and the rest are sample feature data. (2) Perform maximum-minimum normalization on all feature labels to convert them into dimensionless values. The mean of the adjusted input data is 0, and the standard deviation is 1, to eliminate the influence of different dimensions between different parameters and avoid problems such as reduced model error and prolonged training time caused by data with different dimensions during model training. Let the original value of a certain feature be... , The minimum value in the data. If the maximum value in the data is the normalized value, then... The calculation formula is: ; (3) A Multi-Layer Perceptron (MLP) neural network model was selected. MLP is a typical feedforward artificial neural network, suitable for classifying large amounts of data and establishing complex nonlinear mappings. After obtaining a limited set of input variables and corresponding output variables, the MLP network can be used to simulate the relationship between input and output, thereby establishing a prediction method based on the MLP model. Since the reconstruction scheme in this study is a regression problem that maps several input variables to output variables, the MLP neural network model was used.

[0038] MLP models are networks formed by multiple layers of densely connected information processing nodes, called neurons. An MLP network consists of an input layer, one or more hidden layers, and an output layer. The input layer contains nodes for inputting data, and its number corresponds to the number of data variables. The hidden and output layers consist of neurons and output variables, respectively. Adjacent layers are fully connected, while nodes within the same layer are not connected. This study uses four hidden layers: the first three are fully connected layers with 128 neurons each, and the last layer is a fully connected layer with only one neuron.

[0039] The information processing mechanism of MLP neural networks is as follows: ; in, It is the input of the neuron. and These represent the connection weights and biases, respectively. is the activation function for the neuron to capture the nonlinear information of the input. Since the MLP network is applied to a regression problem in this study and the output is arbitrary, the activation function should not be used in the last hidden layer. Therefore, the output mechanism of the output layer is as follows: ; Through the processing and layer-by-layer transmission of information by neurons, the MLP model completes the mapping from input to output. The essence of MLP model learning is to implement the gradient descent algorithm through the forward propagation of information and the backward propagation of error within the network, continuously adjusting the network weight matrix. and bias matrix This minimizes the error between the network's ideal output and the actual output.

[0040] The loss function is set to MSE, the optimizer is Adam, the batch size is 300, and the maximum number of iterations is 500.

[0041] Mean squared error (MSE) is a commonly used error in regression loss functions; it represents the predicted value. With target value The mean of the sum of squares of the differences between them: .

[0042] This invention improves the above training process as follows: Step a, Initialization: Assign initial pheromone concentrations to all parameter combinations in the solution space, randomly generate m ants, and each ant randomly selects nodes (parameter values) in each dimension to form initial parameter combinations; Step b, Node Construction (Parameter Combinations): Each ant selects nodes in each dimension according to probability selection rules; Step c, Fitness Calculation: The parameter combination of each ant is used to construct an MLP model, trained on the training set, and the corresponding loss value is used as the fitness value; Step d, Pheromone Update: Pheromones on all paths evaporate proportionally, and the ant with the best fitness (global optimal solution) releases pheromones on its path; Step e, Iteration Termination: Repeat steps b to d until the maximum number of iterations is reached, and output the globally optimal parameter combination.

[0043] Based on the aforementioned MLP model, this embodiment of the invention provides a specific process for predicting the event level corresponding to a low-temperature chilling injury event using an MLP model based on the characteristic quantities of economic forest fruits, including: Step 6: Train the model and predict risks: (1) Divide the sample set constructed in step 5 into a training sample set (Train) and a test sample set (Test) in an 8:2 ratio, and distinguish between label data (Y) and feature data (X). (2) Substitute the label data (Train_Y) and feature data (Train_X) of the training sample set in (1) into the MLP model in step 5 for iterative training; (3) Substitute the test sample set features (Test_X) from (1) into the model trained in (2) to obtain the predicted value (Test_Y'). Compare Test_Y' and Test_Y to evaluate the trained model. Based on the model characteristics, this scheme uses mean absolute error and root mean square error to evaluate the model accuracy.

[0044] Mean Absolute Error (MAE) measures the average absolute difference between predicted and actual values, representing the average degree of deviation between the predicted and actual values. The calculation formula is as follows: ; in, Represents the total number of samples. and These represent the predicted value and the actual value, respectively.

[0045] The root mean square error (RMSE) is the square root of the ratio of the square of the deviation between the predicted and actual values ​​to the number of observations. The formula is as follows: ; in, Represents the total number of samples. and These represent the predicted value and the actual value, respectively.

[0046] (4) Judge the model quality based on the model evaluation metrics MAE and RMSE. The closer MAE and RMSE are to 0, the higher the model accuracy. Otherwise, repeat step 5 (3) and subsequent steps to further optimize the model parameters and retrain to continuously improve the model training effect.

[0047] Step 7, Model Application – Create a daily low-temperature chilling injury risk warning map for the flowering period of economic forest fruit trees over the next 1-3 days: (1) Based on observations from commercial satellites or drones, determine whether the economic fruit trees in the target area are in bloom; (2) For economic forests and fruit trees and regions that are in the flowering period, extract meteorological satellite surface temperature data, meteorological station data and geographical features, and perform data preprocessing. (3) Input the feature quantities into the optimized model to obtain the cold damage risk level of the economic forest and fruit area, and realize the production of a one-map of the low temperature cold damage risk of economic forest and fruit flowering period in the next 1-3 days.

[0048] The risk level of chilling injury in economic forest and fruit areas is divided into five levels: low risk, relatively low risk, medium risk, relatively high risk, and high risk.

[0049] Table 3. Classification Standards for Cold Damage Risk Assessment in Economic Forest and Fruit Areas

[0050] In summary, this invention, based on meteorological station data, geographic information, planting information of economic forest fruits, flowering periods, etc., utilizes the powerful learning and predictive capabilities of a multi-layer sensing model, combined with advanced remote sensing methods such as drones and commercial satellites, as well as grid point forecast data for temperatures in the next 1-3 days, to conduct early assessment and warning of the risk of low-temperature chilling injury to economic forest fruits during their flowering period. This method comprehensively considers multiple influencing factors, such as meteorological conditions, geographical features, physiological characteristics of economic forest fruits, and historical disaster records. Through data analysis and model training, it achieves accurate prediction and quantitative assessment of low-temperature chilling injury risk, aiming to provide a refined risk map of the fruit flowering period in the next 1-3 days. This helps farmers, agricultural enterprises, and relevant departments to understand the potential risks and impacts of low-temperature chilling injury on economic forest fruits in a timely and accurate manner, thereby taking effective preventive and response measures in advance, reducing economic losses caused by disasters, and ensuring the stable production of economic forest fruits and the sustainable development of the agricultural economy.

[0051] The embodiments of the present invention have at least the following characteristics: (1) Traditional methods for risk assessment and early warning of low-temperature chilling injury in economic forests and orchards mainly include statistical analysis based on historical data, simulation based on physical models, and expert evaluation based on experience. These methods can provide valuable references for risk assessment of low-temperature chilling injury to a certain extent, but they either rely heavily on historical data and are difficult to fully capture the nonlinear relationships in the data, or they have complex parameters, large computational load, and are easily affected by various uncertain factors. The embodiments of this invention leverage the advantages of artificial intelligence, such as its strong generalization ability, better adaptability, and more comprehensive consideration of factors, to construct a risk assessment and early warning model for low-temperature chilling injury in economic forests and orchards. This model can reduce reliance on cumbersome data, fully capture the implicit features in the data, and reduce the uncertainty of prediction.

[0052] (2) The embodiments of the present invention comprehensively consider various influencing factors, such as meteorological conditions, geographical features, physiological characteristics of economic forest fruits and historical disaster records. Through data analysis and model training, combined with temperature grid prediction data, the invention achieves accurate quantitative assessment and early warning of low temperature damage risk. The resulting detailed risk map of low temperature damage during the flowering period of economic forest fruits in the next 1-3 days can provide farmers, agricultural meteorological insurance and related departments with more accurate and reliable prediction and decision-making basis.

[0053] Based on the foregoing embodiments, this invention provides a refined low-temperature chilling injury risk assessment and early warning device for economic forest fruit trees during flowering. (See also...) Figure 3 The schematic diagram shown is of a sophisticated low-temperature chilling injury risk assessment and early warning device for fruit trees during flowering, comprising: Data acquisition module 302 is used to acquire a multi-source historical data set corresponding to the study area. The study area is planted with economic forests and fruits. The multi-source historical data set includes historical single-station low temperature cold damage event data, ground meteorological station observation data, geographical feature data and satellite observation data. The data processing module 304 is used to preprocess the multi-source historical data set to obtain feature quantities that match the fruit growth area and flowering period of economic forests and fruits, and to determine the event level corresponding to the low temperature damage event that occurred in the study area based on the preprocessed historical single-station low temperature damage event data. The model training module 306 is used to train the neural network model based on the characteristic quantities of economic forest fruits and the event level corresponding to the low temperature chilling injury event, so as to obtain the chilling injury risk assessment and early warning model. The chilling injury risk assessment and early warning model is used to assess and warn of the event level corresponding to the low temperature chilling injury event that occurs in the forest fruit growing area during the flowering period.

[0054] The refined low-temperature chilling injury risk assessment and early warning device for economic forest fruits provided in this embodiment of the invention utilizes multi-source data such as historical single-station low-temperature chilling injury event data, ground meteorological station observation data, geographical feature data, and satellite observation data, combined with the growth area and flowering period of economic forest fruits, and leverages the powerful learning and predictive capabilities of neural network models to conduct early assessment and early warning of the risk of low-temperature chilling injury to economic forest fruits during their flowering period.

[0055] In one implementation, the data processing module 304 is specifically used for: Based on the flowering period of economic forest fruits, time matching was performed on historical single-station low-temperature cold damage event data and ground meteorological station observation data in the multi-source historical dataset; based on the forest fruit growth area of ​​economic forest fruits, spatial matching was performed on geographical feature data and satellite observation data in the multi-source historical dataset. Preprocessing was performed on the matched historical single-station low-temperature chilling injury event data, ground meteorological station observation data, geographic feature data, and satellite observation data to obtain characteristic quantities that match the fruit growth area and flowering period of economic forests and fruits. The characteristic quantities include the statistics of historical single-station low-temperature chilling injury event data and meteorological station observation data, geographic feature data, and surface temperature data.

[0056] In one implementation, the data processing module 304 is specifically used for: Based on the preprocessed historical single-station low-temperature chilling injury event data, the characteristic quantities of low-temperature chilling injury events during each fruit and forest flowering period are statistically analyzed. The characteristic quantities of low-temperature chilling injury events include one or more of the following: number of days affected by chilling injury, minimum temperature during chilling injury, average temperature during chilling injury, maximum chilling injury intensity, and average chilling injury intensity. Based on the pre-set cold damage risk index system and the characteristic quantities of low-temperature cold damage events, the event level corresponding to the low-temperature cold damage events occurring in the study area is determined.

[0057] In one implementation, the neural network employs a multilayer perceptron model.

[0058] In one implementation, the model training module 306 is specifically used for: A sample set is constructed based on the characteristics of economic forest fruits. The event level corresponding to the low temperature damage event is used as the sample label. The sample set and its corresponding sample label are divided into a training set and a test set. The training set is used to train the multilayer perceptron model.

[0059] In one implementation, the model training module 306 is specifically used for: Using the test set, the mean absolute error and root mean square error of the cold damage risk assessment and early warning model are determined to evaluate the quality of the model.

[0060] In one implementation, the model training module 306 is specifically used for: Obtain satellite or drone observation results corresponding to the study area to determine whether the economic fruit trees planted in the study area are in the flowering period; If so, based on the fruit-growing area of ​​economic forests, extract the corresponding feature quantities of economic forests from the multi-source current dataset corresponding to the study area; The chilling injury risk assessment and early warning model is used to assess and warn of the event level of low temperature chilling injury events occurring in the fruit growing area based on the characteristic quantities corresponding to economic forests and fruits.

[0061] The device provided in this embodiment of the invention has the same implementation principle and technical effect as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0062] This invention provides an electronic device, specifically, the electronic device includes a processor and a memory; the memory stores a computer program, which, when run by the processor, executes the method described in any of the above embodiments.

[0063] Figure 4 The present invention provides a schematic diagram of the structure of an electronic device 100, which includes a processor 40, a memory 41, a bus 42 and a communication interface 43. The processor 40, the communication interface 43 and the memory 41 are connected through the bus 42. The processor 40 is used to execute executable modules, such as computer programs, stored in the memory 41.

[0064] The memory 41 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 43 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc.

[0065] Bus 42 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0066] The memory 41 is used to store programs. After receiving an execution instruction, the processor 40 executes the program. The method executed by the device for defining the flow process disclosed in any of the foregoing embodiments of the present invention can be applied to the processor 40 or implemented by the processor 40.

[0067] Processor 40 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 40 or by instructions in software form. Processor 40 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 41. The processor 40 reads the information in memory 41 and, in conjunction with its hardware, completes the steps of the above method.

[0068] The computer program product of the readable storage medium provided in the embodiments of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the foregoing method embodiments. For specific implementation, please refer to the foregoing method embodiments, which will not be repeated here.

[0069] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a portion 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.

[0070] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, 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, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A refined method for assessing and warning of low-temperature chilling damage during the flowering period of fruit trees in economic forests, characterized in that, include: A multi-source historical data set corresponding to the study area is obtained. The study area is planted with economic fruit trees. The multi-source historical data set includes historical single-station low temperature cold damage event data, ground meteorological station observation data, geographical feature data and satellite observation data. The multi-source historical data set is preprocessed to obtain feature quantities that match the fruit growth area and flowering period of the economic forest fruit, and the event level corresponding to the low temperature damage event that occurred in the study area is determined based on the preprocessed historical single-station low temperature damage event data. Based on the characteristic quantities of the economic forest fruits and the event level corresponding to the low temperature chilling injury event, a neural network model is trained to obtain a chilling injury risk assessment and early warning model. The chilling injury risk assessment and early warning model is used to assess and warn of the event level corresponding to the low temperature chilling injury event occurring in the forest fruit growing area during the flowering period of the forest fruits.

2. The refined method for assessing and warning of low-temperature chilling injury risk during the flowering period of fruit trees in economic forests according to claim 1, characterized in that, The multi-source historical data set is preprocessed to obtain feature quantities that match the fruit tree growth area and flowering period of the economic fruit trees, including: Based on the flowering period of the economic forest fruits, time matching is performed on the historical single-station low-temperature cold damage event data and the ground meteorological station observation data in the multi-source historical data set; and spatial matching is performed on the geographical feature data and the satellite observation data in the multi-source historical data set based on the forest fruit growth area of ​​the economic forest fruits. The historical single-station low-temperature chilling injury event data, the ground meteorological station observation data, the geographic feature data, and the satellite observation data are preprocessed to obtain feature quantities that match the fruit growth area and flowering period of the economic forest fruit. The feature quantities include the statistics of historical single-station low-temperature chilling injury event data and meteorological station observation data, geographic feature data, and surface temperature data.

3. The refined method for assessing and warning of low-temperature chilling injury risk during the flowering period of economic forest fruit trees according to claim 1, characterized in that, Based on the preprocessed historical single-station low-temperature chilling injury event data, the event levels corresponding to the low-temperature chilling injury events occurring in the study area are determined, including: Based on the preprocessed historical single-station low-temperature chilling injury event data, the characteristic quantities of the low-temperature chilling injury events during each period of the fruit tree flowering season are statistically analyzed. The characteristic quantities of the low-temperature chilling injury events include one or more of the following: number of days affected by chilling injury, minimum temperature during chilling injury, average temperature during chilling injury, maximum chilling injury intensity, and average chilling injury intensity. Based on the preset cold damage risk index system and the characteristic quantities of the low-temperature cold damage events, the event level corresponding to the low-temperature cold damage events occurring in the study area is determined.

4. The refined method for assessing and warning of low-temperature chilling injury risk during the flowering period of economic forest fruit trees according to claim 1, characterized in that, The neural network uses a multilayer perceptron model.

5. The refined method for assessing and warning of low-temperature chilling injury risk during the flowering period of fruit trees in economic forests according to claim 4, characterized in that, Based on the characteristic quantities of the economic forest fruits and the event level corresponding to the low-temperature chilling injury event, the neural network model is trained, including: A sample set is constructed based on the characteristics of the economic forest fruits. The event level corresponding to the low temperature damage event is used as the sample label. The sample set and its corresponding sample label are divided into a training set and a test set. The training set is used to train the multilayer perceptron model.

6. The refined method for assessing and warning of low-temperature chilling injury risk during the flowering period of fruit trees in economic forests according to claim 5, characterized in that, The method further includes: Using the test set, the mean absolute error and root mean square error corresponding to the cold damage risk assessment and early warning model are determined to evaluate the quality of the cold damage risk assessment and early warning model.

7. The refined method for assessing and warning of low-temperature chilling injury risk during the flowering period of fruit trees in economic forests according to claim 1, characterized in that, The method further includes: Obtain satellite or UAV observation results corresponding to the study area to determine whether the economic fruit trees planted in the study area are in the flowering period of the fruit trees; If so, based on the fruit growing area of ​​the economic forest, the feature quantity corresponding to the economic forest is extracted from the multi-source current data set corresponding to the study area; The chilling injury risk assessment and early warning model is used to assess and warn of the event level corresponding to the low-temperature chilling injury event occurring in the fruit orchard growing area based on the characteristic quantities corresponding to the economic fruit orchard.

8. A refined low-temperature chilling injury risk assessment and early warning device for fruit trees during flowering, characterized in that, include: The data acquisition module is used to acquire a multi-source historical data set corresponding to the study area, where economic fruit trees are planted. The multi-source historical data set includes historical single-station low-temperature cold damage event data, ground meteorological station observation data, geographic feature data, and satellite observation data. The data processing module is used to preprocess the multi-source historical data set to obtain feature quantities that match the fruit growth area and flowering period of the economic forest fruit, and to determine the event level corresponding to the low temperature damage event that occurred in the study area based on the preprocessed historical single-station low temperature damage event data. The model training module is used to train a neural network model based on the feature quantities of the economic forest fruit and the event level corresponding to the low temperature chilling injury event, so as to obtain a chilling injury risk assessment and early warning model. The chilling injury risk assessment and early warning model is used to assess and warn of the event level corresponding to the low temperature chilling injury event occurring in the forest fruit growing area during the flowering period of the forest fruit.

9. An electronic device, characterized in that, The method includes a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when invoked and executed by a processor, cause the processor to perform the method according to any one of claims 1 to 7.

Images