Agricultural meteorological data processing system and method thereof

Abstract

The invention discloses an agricultural meteorological data processing system and method, and the method comprises the steps: carrying out the preprocessing of historical growth data and historical meteorological data, and obtaining a historical growth sample data set and a historical meteorological data sample set which affect different growth cycles of a to-be-detected crop; training the historical growth sample data set and the historical meteorological data sample set to obtain a crop meteorological model influencing the growth of the to-be-detected crop; acquiring real-time growth data of crops to be detected and real-time meteorological data of the planting area of the crops to be detected, and inputting the real-time growth data and the real-time meteorological data into the crop meteorological model for analysis to obtain meteorological characteristic data influencing crop growth; preprocessing the meteorological characteristic data to form meteorological time sequence data of the crops to be measured in different growth periods; and predicting the growth risk situation of the to-be-detected crop according to the meteorological time sequence data. According to the invention, the meteorological data of each growth stage of crops can be predicted from details, and the precision and reliability of the growth data in the time period are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural information analysis, and in particular to an agricultural meteorological data processing system and method thereof. Background Art

[0002] The planting and growth of crops are closely related to the climate environment. Agricultural production activities are easily affected by natural meteorological factors, such as light, moisture, temperature or wind, which in turn affect the normal production of agriculture and even lead to reduced agricultural production and income. To this end, by collecting and analyzing the meteorological data of crop growth and the changes in growth data, mastering agricultural meteorological data is important for improving the scientificity and accuracy of agricultural meteorological data, and timely adjusting agricultural activities such as planting, irrigation, fertilization, and pest control to maximize crop yield and quality. At present, there are various redundant data in the processing and analysis of agricultural meteorological data, which leads to low agricultural meteorological accuracy and cannot provide more detailed and accurate agricultural meteorological data. Summary of the invention

[0003] The purpose of the present invention is to provide an agricultural meteorological data processing system and method thereof, which can predict the meteorological data of each stage of crop growth in detail, improve the accuracy and reliability of growth data within a time period, reduce the generation of redundant data, and improve the processing efficiency of agricultural meteorological data. In order to achieve the above purpose, the invention adopts the following technical effects:

[0004] According to one aspect of the present invention, the present invention provides an agricultural meteorological data processing method, the agricultural data processing method comprising the following steps:

[0005] Step 1, obtaining historical growth data and historical meteorological data of the crop to be tested, and preprocessing the historical growth data and historical meteorological data to obtain a historical growth sample data set and a historical meteorological data sample set that affect different growth cycles of the crop to be tested;

[0006] Step 2, training the historical growth sample data set and the historical meteorological data sample set to obtain a crop meteorological model that affects the growth of the tested crops;

[0007] Step 3, obtaining real-time growth data of the crop to be tested and real-time meteorological data of the planting area of ​​the crop to be tested, inputting them into the crop meteorological model for analysis, and obtaining meteorological characteristic data that affects the growth of the crop;

[0008] Step 4, preprocessing the meteorological characteristic data to form meteorological time series data of different growth periods of the tested crops;

[0009] Step 5: predict the growth risk situation of the tested crops according to the meteorological time series data, and output corresponding farming activity adjustment suggestions according to the growth risk situation.

[0010] Preferably, in the above scheme, in step 2, training the historical growth sample data set and the historical meteorological data sample set mainly includes:

[0011] Acquire growth parameter sample data and meteorological time series data sample data from historical growth data affecting different growth cycles of the tested crop and historical meteorological data of the tested crop in the growth cycle respectively;

[0012] Several sample points are randomly selected from the growth parameter sample data and the meteorological time series data sample data as training sample sets. The training sample sets are trained using a multi-layer LSTM network to obtain a crop meteorological model.

[0013] Preferably, the above scheme randomly extracts a number of sample points as training sample sets, including: randomly selecting sample point data within 3-5 time periods from the growth parameter sample data and the meteorological time series data sample data in each time segment of each growth cycle, and aggregating the selected sample point data into a training sample set.

[0014] Preferably, in the above scheme, in step 3, inputting the real-time growth data and the real-time meteorological data into the crop meteorological model for analysis comprises the following steps:

[0015] The growth cycle of the crop to be tested is divided into n time nodes, where n is greater than or equal to 2, and k growth characteristic data are randomly selected in each time node, where k is greater than or equal to 20;

[0016] The meteorological data corresponding to k growth characteristic data are extracted, and the growth characteristic data are fused with the meteorological data as input data, which are input into the crop meteorological model for iterative optimization training to obtain meteorological characteristic data.

[0017] Preferably, in the above scheme, the fusion of growth characteristic data and meteorological data includes the following manner: the meteorological data corresponding to each growth characteristic data is sequentially spliced ​​according to the time dimension or inserted and spliced ​​into the previous time dimension.

[0018] In the above scheme, the step 4 of preprocessing the meteorological characteristic data to form meteorological time series data of different growth periods of the tested crop comprises the following steps: performing missing smoothing processing on the meteorological characteristic data and eliminating abnormal data points in the characteristic data;

[0019] The real-time growth data of each time node in different growth cycles are labeled with corresponding meteorological characteristic data and integrated in time series to construct time series sample growth data of the crops to be tested, and the key meteorological factors affecting the growth of the crops to be tested are selected based on the time series sample growth data;

[0020] The key meteorological factors are classified in parallel, and the meteorological time series data of the classified key meteorological factors are output in chronological order.

[0021] According to another aspect of the present invention, the present invention provides an agricultural meteorological data processing system, which includes a data collection module, a model construction module, a real-time data collection module, a data fusion module, a data processing module and a data prediction module; the data collection module is used to collect historical growth data and historical meteorological data of crops to be tested from multiple channels; the model construction module is used to analyze and train historical data to form a crop meteorological model; the real-time data collection module is used to collect real-time growth data and real-time meteorological data of the planting area of ​​the crops to be tested in real time; the data fusion module is used to fuse the collected real-time growth data and real-time meteorological data, and input the fused data into the crop meteorological model for iterative optimization training to obtain meteorological characteristic data; the data processing module is used to process the meteorological characteristic data, and construct the time series sample growth data of the crops to be tested with the processed data, and extract key meteorological factors in chronological order based on the time series sample growth data; the data prediction module predicts the growth risk situation of the crops to be tested and generates agricultural activity adjustment suggestions based on the key meteorological factors.

[0022] Preferably, the above scheme further comprises a data classification module and a data splitting and splicing module; the data classification module is used to classify and organize key meteorological factors, and generate meteorological time series data according to chronological order for the key meteorological factors; the data prediction module is based on predicting the growth risk situation of the tested crops and generating suggestions for adjusting farming activities; the data splitting and splicing module is used to split the real-time growth data and real-time meteorological data for training, and splice and fuse the growth characteristic data after training with the meteorological data.

[0023] The invention adopts the above technical solution, and the invention has the following technical effects:

[0024] (1) The present invention can predict the growth risk of the tested crops by obtaining key meteorological factors that have a significant impact on the growth of crops, so as to make agricultural adjustments to the growth of crops at various time points, improve the growth environment of crops, and achieve the purpose of improving the growth quality and yield of crops. It can thus make corresponding field management decisions for the growth of crops, effectively improve the efficiency of crop production activities, improve crop quality, and establish an efficient and high-quality agricultural production management model.

[0025] (2) The present invention can predict the meteorological data of each stage of crop growth in detail, improve the accuracy and reliability of growth data within a time period, reduce the generation of redundant data, improve the processing efficiency of agricultural meteorological data, and improve the reliability of predicting the growth risk of the crops to be tested. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a flowchart of an invented agricultural meteorological data processing method;

[0027] Figure 2 It is a schematic diagram of the system principle model of an invented agricultural meteorological data processing system; DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the invention more clearly understood, the invention is further described in detail below with reference to the accompanying drawings and preferred embodiments. However, it should be noted that many details listed in the specification are only for the purpose of enabling the reader to have a thorough understanding of one or more aspects of the invention, and these aspects of the invention can be realized even without these specific details.

[0029] like Figure 1 As shown, according to the invention, a method for processing agricultural meteorological data is provided, and the agricultural data processing method comprises the following steps:

[0030] Step 1, obtain historical growth data and historical meteorological data of the crops to be tested, and pre-process the historical growth data and historical meteorological data to obtain historical growth sample data sets and historical meteorological data sample sets that affect different growth cycles of the crops to be tested; the crops to be tested are various planted crops, and the historical meteorological data of the planted crops in the growth cycle of the planting area generally includes the meteorological data including wind speed and direction data, forecast rainfall, rainfall data (including daily rainfall, current rainfall and accumulated rainfall), air temperature and humidity data (average temperature, average maximum temperature, average minimum temperature, average relative humidity, etc.), ground temperature, sunshine data (daily The test crops take the growth cycles of grain crops (such as rice, wheat, corn, soybean, etc.) such as seedling, growth, flowering and maturity as an example. The historical growth data of the test crops in the planting area include crop variety information, growth latitude data, and soil parameters, water quality parameters, pest and disease information, branch and leaf growth images or videos, root and stem growth data, fertilization information, and multiple growth element data such as crop growth remote sensing data in each time period of the growth cycle (seedling period, sowing period, seedling period, growth period, flowering period, fruiting period and maturity period);

[0031] Step 2, training the historical growth sample data set and the historical meteorological data sample set to obtain a crop meteorological model that affects the growth of the crops to be tested; in the present invention, training the historical growth sample data set and the historical meteorological data sample set mainly includes: first, acquiring growth parameter sample data and meteorological time series data sample data from the historical growth data that affects the crops to be tested in different growth cycles and the historical meteorological data of the crops to be tested in the growth cycle; second, randomly selecting a number of sample points from the growth parameter sample data and the meteorological time series data sample data as training sample sets, and training the training sample sets using a multi-layer LSTM network to obtain a crop meteorological model; wherein, randomly selecting a number of sample points as training sample sets includes: randomly selecting sample point data within 3-5 time periods from the growth parameter sample data and the meteorological time series data sample data in each time segment of each growth cycle, and aggregating the selected sample point data into a training sample set

[0032] Step 3, obtaining the real-time growth data of the crop to be tested and the real-time meteorological data of the planting area of ​​the crop to be tested, inputting them into the crop meteorological model for analysis, and obtaining the meteorological characteristic data that affects the growth of the crop; in the present invention, obtaining the real-time growth data of the crop to be tested and the real-time meteorological data of the planting area of ​​the crop to be tested is to collect the required monitoring data in real time through remote sensing equipment, Internet of Things equipment (various sensors, location positioning modules, image or video sensors); wherein, inputting the real-time growth data and the real-time meteorological data into the crop meteorological model for analysis includes the following steps:

[0033] Step 31, dividing the growth period of the crop to be tested into n time nodes, n is greater than or equal to 2, and randomly selecting k growth characteristic data in each time node, k is greater than or equal to 20;

[0034] Step 32, extracting meteorological data corresponding to k growth characteristic data, fusing the growth characteristic data with the meteorological data as input data, inputting the crop meteorological model for iterative optimization training, and obtaining meteorological characteristic data.

[0035] In the present invention, after the data set of the historical growth of the tested crop is correspondingly associated with its corresponding historical meteorological data, a historical data sample set (time series data sample data) with a time series is formed; part of the data is randomly extracted from the historical data sample set as a training sample set, the training sample set is trained and verified, and the trained crop meteorological model is adjusted by increasing the training iteration time, increasing the number of training sample sets, and optimizing the loss function and the deep learning network model, and the crop meteorological model with excellent results is selected as the data analysis model for analyzing the real-time growth data and real-time meteorological data of the tested crop; meteorological characteristic data of crop growth is established, so that corresponding field management decisions can be made for crop growth, the efficiency of crop production activities is effectively improved, the quality of crops is improved, and an efficient and high-quality agricultural production management model is established; the present invention uses a trained crop meteorological model based on a multi-layer LSTM network to analyze the real-time growth data and real-time meteorological data changes of the growing week, so as to obtain the trend of agricultural growth changes and the meteorological factors affecting the growth changes, and make accurate long-term predictions, which can provide a reliable prediction model and effectively guide the formulation of agricultural production activities. The multi-layer LSTM network includes an input layer, an LSTM layer, a hidden layer, and an output layer. For the trained crop meteorological model based on the LSTM network, the time-series growth data and real-time meteorological data are input into the input layer after normalization processing, and then the key meteorological factors corresponding to each other in the time series data are extracted through the LSTM layer. The hidden layer processes the key meteorological factors corresponding to each other in the growth cycle, and outputs the final predicted feature data through output layer mapping.

[0036] Step 4, preprocessing the meteorological characteristic data to form meteorological time series data of different growth periods of the tested crops; wherein forming the meteorological time series data of different growth periods of the tested crops includes the following steps;

[0037] Step 41, performing missing smoothing processing on the meteorological characteristic data and removing abnormal data points in the characteristic data; processing the meteorological characteristic data to reduce data noise and reduce data deviation;

[0038] Step 42, annotating the real-time growth data of each time node in different growth cycles with corresponding meteorological characteristic data and fusing them in time series to construct time series sample growth data of the crop to be tested, and selecting key meteorological factors affecting the growth of the crop to be tested based on the time series sample growth data;

[0039] Step 43, classify the key meteorological factors in parallel, and output meteorological time series data for the classified key meteorological factors in chronological order;

[0040] Step 5, predicting the growth risk situation of the crops to be tested according to the meteorological time series data, and outputting corresponding agricultural activity adjustment suggestions according to the growth risk situation; in the present invention, by fusing the real-time growth data of different time nodes with the corresponding meteorological characteristic data, when fusing the growth characteristic data in the planting area with the meteorological data, the meteorological data corresponding to each growth characteristic data is sequentially spliced ​​according to the time dimension (timestamp), or the meteorological data is moved forward or backward by one time node in the current time dimension (timestamp) position for insertion and splicing, so as to form growth characteristic data and meteorological data that are interrelated under the same longitude and latitude and different time dimensions, so as to obtain growth prediction data under different time dimensions, thereby overcoming the unpredictable occurrence of abnormal meteorological weather during crop growth, being able to fully understand the growth situation and meteorological conditions of crops, and improving the accuracy of meteorological data on crop growth. Obtaining time-series sample growth data with timestamps, it is possible to intuitively read the time-series sample growth data of the tested crops at each time point, and then select various key meteorological factor characteristic data that affect crop growth from the time-series sample growth data. By analyzing different types of key meteorological factors (data) that affect crop growth, meteorological time-series data of various key meteorological factors (data) in various time periods can be obtained. For example, soil parameters and meteorological condition data are used to analyze the impact of crop growth characteristic data sets. For crops at various time nodes, taking the flowering period as an example, soil parameters (nutrient content, organic matter content, soil temperature and humidity) are analyzed. , pH value) analysis, comprehensively consider meteorological factors such as air temperature and humidity, predicted rainfall, rainfall data, sunshine data, wind speed and direction, atmospheric pressure and total emissivity corresponding to the current time node, analyze various meteorological factors at each time node respectively, and obtain the influence of meteorological factors on soil parameters of crop growth, that is, obtain soil parameter-meteorological time series data of crop flowering stage, and identify key meteorological factors that affect large changes in soil parameters, so as to predict the growth risk of the tested crops, make agricultural adjustments to the growth of crops at this time node, improve the growth environment of crops, and achieve the purpose of improving crop growth quality and yield.

[0041] According to another aspect of the present invention, Figure 2The present invention provides an agricultural meteorological data processing system, wherein the processing system comprises a data collection module, a model construction module, a real-time data collection module, a data fusion module, a data processing module and a data prediction module; the data collection module is used to collect historical growth data and historical meteorological data of crops to be tested from multiple channels; the model construction module is used to analyze and train historical data to form a crop meteorological model; the real-time data collection module is used to collect real-time growth data and real-time meteorological data of the planting area of ​​the crops to be tested in real time; the data fusion module is used to fuse the collected real-time growth data and real-time meteorological data, and input the fused data into the crop meteorological model for iterative optimization training to obtain meteorological characteristic data; the data processing module is used to process the meteorological characteristic data, and construct the time series sample growth data of the crops to be tested with the processed data, and extract key meteorological factors in chronological order based on the time series sample growth data; the data prediction module predicts the growth risk situation of the crops to be tested and generates agricultural activity adjustment suggestions based on the key meteorological factors. By predicting the key meteorological factors that have a significant impact on crop growth, the reliability of the growth risk prediction of the crops to be tested is improved, so that agricultural adjustments can be made to crop growth at that time point, improving the crop growth environment and achieving the goal of improving crop growth quality and yield.

[0042] In the present invention, the processing system further includes a data classification module and a data splitting and splicing module; the data classification module is used to classify and organize key meteorological factors, and generate meteorological time series data according to the time sequence of key meteorological factors; the data prediction module is based on predicting the growth risk situation of the tested crops and generating suggestions for adjusting agricultural activities; the data splitting and splicing module is used to split the real-time growth data and real-time meteorological data for training, splice and fuse the growth characteristic data after training with the meteorological data, split the real-time growth data and real-time meteorological data for training according to each time period, obtain the training data of each time period, and after the training is completed, fuse and splice the real-time growth data of different time nodes with the corresponding meteorological characteristic data to form interrelated growth characteristic data and meteorological data under each time dimension, that is, input the sample data set for training in the crop meteorological model. Thus, the meteorological data of each stage of crop growth can be predicted in detail, the accuracy and reliability of the growth data in the time period can be improved, and the generation of redundant data can be reduced.

[0043] The above is only a preferred embodiment of the invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the invention. These improvements and modifications should also be regarded as the scope of protection of the invention.

Claims

1. A method for processing agricultural meteorological data, characterized in that: The agricultural data processing method comprises the following steps: Step 1, obtaining historical growth data and historical meteorological data of the crop to be tested, and preprocessing the historical growth data and historical meteorological data to obtain a historical growth sample data set and a historical meteorological data sample set that affect different growth cycles of the crop to be tested; Step 2, training the historical growth sample data set and the historical meteorological data sample set to obtain a crop meteorological model that affects the growth of the tested crops; Step 3, obtaining real-time growth data of the crop to be tested and real-time meteorological data of the planting area of ​​the crop to be tested, inputting them into the crop meteorological model for analysis, and obtaining meteorological characteristic data that affects the growth of the crop; Step 4, preprocessing the meteorological characteristic data to form meteorological time series data of different growth periods of the tested crops; Step 5: predict the growth risk situation of the tested crops based on the meteorological time series data, and output corresponding farming activity adjustment suggestions based on the growth risk situation.

2. The method for processing agricultural meteorological data according to claim 1, characterized in that: In step 2, training the historical growth sample data set and the historical meteorological data sample set mainly includes: Acquire growth parameter sample data and meteorological time series data sample data from historical growth data affecting different growth cycles of the tested crop and historical meteorological data of the tested crop in the growth cycle respectively; Several sample points are randomly selected from the growth parameter sample data and the meteorological time series data sample data as training sample sets. The training sample sets are trained using a multi-layer LSTM network to obtain a crop meteorological model.

3. The method for processing agricultural meteorological data according to claim 2, characterized in that: Randomly selecting a number of sample points as training sample sets includes: randomly selecting sample point data within 3-5 time periods from growth parameter sample data and meteorological time series data sample data in each time segment of each growth cycle, and aggregating the selected sample point data into a training sample set.

4. The agricultural meteorological data processing method according to claim 1, characterized in that: In step 3, inputting the real-time growth data and the real-time meteorological data into the crop meteorological model for analysis comprises the following steps: The growth cycle of the crop to be tested is divided into n time nodes, where n is greater than or equal to 2, and k growth characteristic data are randomly selected in each time node, where k is greater than or equal to 20; The meteorological data corresponding to k growth characteristic data are extracted, the growth characteristic data are fused with the meteorological data as input data, and the crop meteorological model is input for iterative optimization training to obtain meteorological characteristic data.

5. The agricultural meteorological data processing method according to claim 4, characterized in that: The growth characteristic data and the meteorological data are integrated in the following manners: the meteorological data corresponding to each growth characteristic data is sequentially spliced ​​according to the time dimension or inserted and spliced ​​into the previous time dimension.

6. The method for processing agricultural meteorological data according to claim 1, characterized in that: The step 4, pre-processing the meteorological characteristic data to form meteorological time series data of different growth periods of the tested crops, includes the following steps: performing missing smoothing processing on the meteorological characteristic data and eliminating abnormal data points in the characteristic data; The real-time growth data of each time node in different growth cycles are labeled with corresponding meteorological characteristic data and integrated in time series to construct time series sample growth data of the crops to be tested, and the key meteorological factors affecting the growth of the crops to be tested are selected based on the time series sample growth data; The key meteorological factors are classified in parallel, and the meteorological time series data of the classified key meteorological factors are output in chronological order.

7. The agricultural meteorological data processing method according to claim 1, characterized in that: The growth data information of the crops to be predicted includes but is not limited to crop variety information, seedling period, sowing period, seedling period, growth period, flowering period, fruiting period and maturity period; the meteorological data includes wind speed and direction data, predicted rainfall, rainfall data, air temperature and humidity data, ground temperature, sunshine data, atmospheric pressure and total radiation data.

8. An agricultural meteorological data processing system, characterized in that: The processing system includes a data collection module, a model building module, a real-time data acquisition module, a data fusion module, a data processing module and a data prediction module; the data collection module is used to collect historical growth data and historical meteorological data of the tested crops from multiple channels; the model building module is used to analyze and train the historical data to form a crop meteorological model; The real-time data acquisition module is used to collect real-time growth data and real-time meteorological data of the crop planting area to be tested; the data fusion module is used to fuse the collected real-time growth data and real-time meteorological data, and input the fused data into the crop meteorological model for iterative optimization training to obtain meteorological characteristic data; the data processing module is used to process the meteorological characteristic data, and construct the time series sample growth data of the crop to be tested with the processed data, and extract key meteorological factors in chronological order based on the time series sample growth data; the data prediction module predicts the growth risk situation of the crop to be tested according to the key meteorological factors and generates suggestions for adjusting agricultural activities.

9. An agricultural meteorological data processing system according to claim 8, characterized in that: The processing system also includes a data classification module and a data splitting and splicing module; the data classification module is used to classify and organize key meteorological factors, and generate meteorological time series data according to chronological order; the data prediction module is based on predicting the growth risk situation of the tested crops and generating suggestions for adjusting agricultural activities; the data splitting and splicing module is used to split the real-time growth data and real-time meteorological data for training, and splice and fuse the growth characteristic data after training with the meteorological data.

Images