Intelligent and efficient water-saving irrigation system for paddy fields

Abstract

The present invention relates to the technical field of water conservancy irrigation, and discloses an intelligent and efficient water-saving irrigation system for paddy fields. The system includes a multi-source data fusion module, an irrigation prediction module, an efficient water-saving irrigation module, and an early warning and monitoring module. First, crop irrigation data is collected, and the multi-source data is fused to generate a crop irrigation joint data set. The adaptive irrigation threshold is calculated. When the soil moisture content is greater than the adaptive irrigation threshold, the irrigation time and irrigation amount are predicted based on a neural network. When the soil moisture content is less than or greater than the adaptive irrigation threshold, a staged control strategy is adopted for staged irrigation, and then the crops are precisely irrigated in combination with a spatiotemporal precision irrigation method. Finally, a dry season-rainy season intelligent switching strategy is used for dynamic water storage, and real-time early warning and remote monitoring of the water level are performed. The present invention achieves the purpose of intelligent and efficient water-saving irrigation for paddy fields by performing efficient water-saving irrigation on paddy fields, and the method is objective and accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of water conservancy irrigation, and in particular to an intelligent and efficient water-saving irrigation system for paddy fields. Background Art

[0002] Global water shortages are becoming increasingly severe. Agricultural irrigation accounts for over 70% of total freshwater consumption. Rice cultivation, as the most water-intensive crop, directly impacts the sustainable use of water resources. Improving water-saving efficiency in paddy field irrigation has become a core issue for ensuring food security and ecological balance.

[0003] Traditional water-saving irrigation methods for paddy fields mainly adopt a combination of engineering water conservation and agronomic water conservation, including three typical modes: intermittent irrigation, shallow wet irrigation, and controlled irrigation. Among them, intermittent irrigation reduces ineffective leakage by periodically draining and drying the fields; shallow wet irrigation uses alternating "shallow water layer-wet water layer-shallow water layer" water layer management; and controlled irrigation sets the lower limit of soil moisture control based on the crop water requirement threshold.

[0004] Traditional water-saving irrigation methods for paddy fields rely on manual experience to set fixed irrigation cycles and water volume parameters. Although they can reduce water consumption to a certain extent, they lack dynamic regulation capabilities and are prone to causing drought and flood stress or excessive irrigation, resulting in large fluctuations in water-saving effects and poor adaptability in extreme weather conditions. Summary of the Invention

[0005] In response to the problems in the related art, the present invention provides an intelligent and efficient water-saving irrigation system for paddy fields to overcome the technical problems existing in the existing related art.

[0006] To solve the above technical problems, the present invention is achieved through the following technical solutions:

[0007] The present invention is an intelligent and efficient water-saving irrigation system for paddy fields, which specifically includes: a multi-source data fusion module, an irrigation prediction module, a high-efficiency water-saving irrigation module, and an early warning and monitoring module;

[0008] The multi-source data fusion module is used to collect data on the dynamic changes of crop growth indicators, soil moisture conditions and meteorological factors, and generate a joint crop irrigation data set using multi-source data fusion;

[0009] The irrigation prediction module is used to calculate the adaptive irrigation threshold, compare the soil moisture content with the adaptive irrigation threshold, and when the soil moisture content is greater than the adaptive irrigation threshold, predict the irrigation time and irrigation amount based on the neural network to obtain an irrigation prediction value;

[0010] The efficient water-saving irrigation module is used to calculate the difference in soil irrigation amount based on the irrigation prediction value and the current irrigation value. When the soil moisture content is less than or greater than the adaptive irrigation threshold, a staged control strategy is adopted to carry out staged irrigation, and then a spatiotemporal precision irrigation method is combined to precisely irrigate crops.

[0011] The early warning and monitoring module is used to consider the difference in soil irrigation amount after precise irrigation, use the dry season-rainy season intelligent switching strategy to dynamically store water, and provide real-time early warning and remote monitoring of the water level.

[0012] Preferably, the data collected on the dynamic changes of crop growth indicators, soil moisture conditions and meteorological factors include:

[0013] Selecting a crop planting area, measuring the leaf area of ​​the crops in the crop planting area, and calculating the leaf area index of the crops;

[0014] Set a set of crop leaf area indices, find the crop leaf area index corresponding to the crop leaf area index, obtain the crop growth stage corresponding to the crop leaf area index, and group the same and adjacent crop growth stages in the crop planting area into the same crop planting sub-area, thereby obtaining several identical crop planting sub-areas;

[0015] The soil moisture in the same crop planting sub-area is measured to obtain a crop soil moisture set; then the meteorological data and time characteristics of the crop planting area in previous years are obtained to form a meteorological data set of the crop planting area in previous years and a time characteristic set of the crop planting area in previous years.

[0016] Preferably, the generating of a crop irrigation joint data set by fusing multi-source data comprises:

[0017] The same crop planting sub-region is recorded as a crop planting sub-region set, and is combined with the crop soil moisture set and the crop planting region meteorological data set of previous years to form a crop irrigation joint data set.

[0018] Preferably, the calculating of the adaptive irrigation threshold comprises:

[0019] The water content of crops when they wilt and the maximum water holding capacity of crops in the crop planting sub-area are measured. According to the crop growth stage corresponding to the crop planting sub-area, a dynamic weight coefficient is set, and then the adaptive irrigation threshold corresponding to the crop growth stage is calculated.

[0020] Preferably, when the soil moisture content is greater than the adaptive irrigation threshold, predicting the irrigation time and irrigation amount based on the neural network includes:

[0021] According to the crop irrigation joint data set, the soil moisture content in the crop planting sub-area set is calculated. For different crop planting sub-areas, the soil moisture content and the adaptive irrigation threshold are compared. When the soil moisture content is greater than the adaptive irrigation threshold, the meteorological data set of the crop planting area in previous years is normalized, and a sliding window is set. The meteorological data of the crop planting area in previous years in the sliding window are used to generate rainfall samples to form a rainfall sample set; the time feature set of the crop planting area in previous years is normalized according to the step size. Divide into rainfall time sets;

[0022] The BiLSTM is set to include an input layer, a forward LSTM, a backward LSTM, and an output layer. After the Attention mechanism is added to the backward LSTM, and the pooling layer is added to the Attention mechanism, the SVM classification model is added to the pooling layer, and the output layer is after the pooling layer to construct a hybrid prediction model. The rainfall sample set is divided into a sample training set and a sample test set, and the rainfall time set is divided to obtain a sample training time set and a sample test time set. The sample training set and the sample training time set are input into the hybrid prediction model. The Attention mechanism performs feature learning on the temperature, air pressure, wind speed, humidity, and evaporation, generates time step weights, and aggregates them. The SVM classification model outputs whether it will rain in the next 30 days, using 1 to indicate rainfall and 0 to indicate no rainfall. When rainfall occurs, the output layer outputs the rainfall prediction result, and when no rainfall occurs, the output layer outputs 0. This is continued until the hybrid prediction model converges to obtain a trained hybrid prediction model.

[0023] Then, the sample test set and the sample test time set are input into the trained hybrid prediction model. The accuracy threshold is set. When the accuracy of the output rainfall prediction result is greater than the accuracy threshold, the iteration is stopped to obtain the SVM-Attention-BiLSTM hybrid prediction model. Otherwise, the weight is adjusted until the accuracy of the rainfall prediction result is greater than the accuracy threshold.

[0024] Get the current time The temperature, pressure, wind speed, humidity, evaporation and the corresponding season, month and week number of the month are obtained to obtain the current meteorological data set and the current time feature set. The current meteorological data set is normalized and the current time feature set is processed by sliding window and input into

[0025] In the SVM-Attention-BiLSTM hybrid prediction model, the rainfall prediction value and rainfall prediction time are output;

[0026] The irrigation amount at the current time is obtained again, and the rainfall forecast value within the range from the current time to the rainfall forecast time is counted, and the irrigation amount forecast value and the irrigation time forecast value are calculated and recorded as the irrigation forecast value.

[0027] Preferably, the calculating of the soil irrigation amount difference and adopting a staged control strategy for staged irrigation when the soil moisture content is less than or greater than the adaptive irrigation threshold comprises:

[0028] The irrigation prediction value includes a number of irrigation time prediction values ​​and corresponding irrigation amount prediction values, finding the rated irrigation amount from the current time to the irrigation time prediction value, calculating the difference between the rated irrigation amount and the irrigation amount prediction value, and recording it as the soil irrigation amount difference;

[0029] When the soil moisture content is less than or greater than the adaptive irrigation threshold, determining the crop growth stage corresponding to the set of crop planting sub-areas for different crop planting sub-areas, and then calculating the water requirements of the different crop growth stages; calculating the soil moisture content of the different crop planting sub-areas based on the crop irrigation joint data set;

[0030] The irrigation amount at different crop growth stages is calculated in sequence to obtain the irrigation amount for different crop planting sub-areas, and irrigation equipment is used for irrigation.

[0031] Preferably, the precise irrigation of crops in combination with the spatiotemporal precise irrigation method comprises:

[0032] During the irrigation process, a small amount of multiple irrigation strategy is used as a time-precision irrigation method. The irrigation amount of different crop planting sub-areas is divided according to the maximum water holding capacity of the crops in the crop planting sub-areas to obtain the irrigation frequency. The intermittent time is set, and during the irrigation process using the irrigation equipment, the crop planting sub-areas are irrigated at the maximum water holding capacity of the crops during each intermittent time, and the water level of the crop planting sub-area is recorded.

[0033] A spatial precision irrigation method is obtained by taking the slope into consideration, calculating the slope of the crop planting sub-area according to the water level of the crop planting sub-area, setting the slope level and the corresponding correction coefficient, and obtaining the slope position of the crop planting sub-area;

[0034] The slope position includes a flat slope, a slope top and a slope bottom;

[0035] Set the upper and lower water level thresholds. When the crop planting sub-area is on a flat slope, irrigation will be stopped until the irrigation times are reached or the water level in the crop planting sub-area is greater than the lower water level threshold and less than the upper water level threshold.

[0036] When the crop planting sub-area is at the top of the slope, reduce the number of irrigations or reduce the lower and upper limits of the water level threshold; when the crop planting sub-area is at the bottom of the slope, increase the number of irrigations or increase the lower and upper limits of the water level threshold.

[0037] Preferably, the use of the dry season-rainy season intelligent switching strategy for dynamic water storage includes:

[0038] After precision irrigation, the remaining water volume in the crop planting area is obtained and a water volume threshold is set. When the remaining water volume is less than the water volume threshold, the dry season strategy is introduced. The crop growth stage priority is set. At this time, the crop planting sub-areas are irrigated in descending order according to the crop growth stage priority, and the irrigation volume of the crop planting sub-areas is added to the soil irrigation volume difference. The return water is collected, sedimented, filtered and disinfected, and then re-injected into the reservoir.

[0039] When the remaining water volume is greater than or equal to the water volume threshold, the crop planting area enters the rainy season; the irrigation volume of the crop planting sub-area is subtracted from the soil irrigation volume difference, and the crop planting sub-areas are irrigated in turn; a time threshold and a water holding capacity threshold are set, and when the maximum water holding capacity of the crop is greater than the water holding capacity threshold and the time is maintained for more than the time threshold, the drainage pump is started to drain the water into the reservoir.

[0040] Preferably, the real-time early warning and remote monitoring of the water level includes:

[0041] The water level of the crop planting sub-area is monitored in real time. When the water level of the crop planting sub-area is less than or equal to the lower water level threshold or greater than or equal to the upper water level threshold, an automatic alarm is sounded and the position of the crop planting sub-area is located.

[0042] The present invention has the following beneficial effects:

[0043] 1. This invention collects data on the dynamic changes of crop growth indicators, soil moisture conditions and meteorological factors, and builds a system perception layer to get rid of the reliance on a single soil moisture condition to determine irrigation needs. It focuses on the dynamic correlation between meteorological changes and crop growth stages, and uses multi-source data fusion to lay the foundation for future efficient water-saving irrigation of paddy fields.

[0044] 2. This invention indirectly obtains irrigation prediction values ​​by using a hybrid prediction model, integrating the SVM classification model, attention mechanism and BiLSTM neural network to identify the weights of various meteorological factors, greatly improving the stability and accuracy of the prediction. It is suitable for the precise prediction of extreme rainfall events and greatly increases adaptability in extreme weather. At the same time, the accurate prediction results are helpful for the subsequent calculation of soil irrigation volume differences and irrigation volume correction to achieve better irrigation and water-saving effects.

[0045] 3. The invention adopts a phased control strategy to carry out phased irrigation, that is, irrigation is carried out in different areas according to the different water requirements and rainfall in different growth stages of crops. While ensuring the normal growth of crops, the irrigation volume is reduced, thereby improving the water-saving rate. The invention then combines the time-space precision irrigation method to carry out precise irrigation of crops. Through precise time-space water distribution, gravity water loss is reduced, and irrigation uniformity is significantly improved, thereby ensuring increased yield while achieving efficient water-saving irrigation.

[0046] 4. This invention corrects the irrigation amount by considering the difference in soil irrigation amount, uses the intelligent switching strategy between dry season and rainy season to dynamically store water, and then conducts real-time early warning and remote monitoring of the water level. This method gives priority to high-priority areas, and at the same time implements intelligent linkage of drainage and water storage, so as to achieve control in wet times and accurate irrigation in dry times, and stable water-saving effect.

[0047] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying any creative work.

[0049] Figure 1 The present invention provides a flow chart of an intelligent and efficient water-saving irrigation system for paddy fields.

[0050] Figure 2 The present invention provides a flow chart of an intelligent and efficient water-saving irrigation method for paddy fields. DETAILED DESCRIPTION

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0052] Traditional water-saving irrigation methods for paddy fields rely on manual experience to set fixed irrigation cycles and water volume parameters. Although they can reduce water consumption to a certain extent, they lack dynamic regulation capabilities and are prone to causing drought and flood stress or excessive irrigation, resulting in large fluctuations in water-saving effects and poor adaptability in extreme weather conditions.

[0053] In order to solve the above technical problems, Figure 1As shown, an embodiment of the present invention provides an intelligent and efficient water-saving irrigation system for paddy fields, specifically comprising: a multi-source data fusion module, an irrigation prediction module, an efficient water-saving irrigation module, and an early warning and monitoring module. The multi-source data fusion module is used to collect data on the dynamic changes of crop growth indicators, soil moisture conditions, and meteorological factors, and use multi-source data fusion to generate a joint crop irrigation data set. The irrigation prediction module is used to calculate an adaptive irrigation threshold, compare soil moisture content with the adaptive irrigation threshold, and when the soil moisture content is greater than the adaptive irrigation threshold, predict the irrigation time and irrigation amount based on a neural network to obtain an irrigation prediction value. The efficient water-saving irrigation module is used to calculate the difference in soil irrigation amount based on the irrigation prediction value and the current irrigation value. When the soil moisture content is less than or greater than the adaptive irrigation threshold, a staged control strategy is adopted for staged irrigation, and then a spatiotemporal precision irrigation method is used to precisely irrigate crops. The early warning and monitoring module is used to consider the difference in soil irrigation amount after precision irrigation, dynamically store water using an intelligent dry season-rainy season switching strategy, and provide real-time early warning and remote monitoring of the water level.

[0054] A specific embodiment is that an irrigation system can be formed using components such as a main pipe (agricultural pipe), a branch pipe (capillary pipe), a tee / cross, a water gauge and water level sensor, a flow meter, an intelligent butterfly valve, a mobile soil moisture recorder, and an integrated control box; wherein the main pipe is laid along the ridge, the branch pipe is connected through a tee / cross, and is controlled by an intelligent butterfly valve to serve as an irrigation pipe network unit; a water gauge and water level sensor is used to measure the water level, a flow meter records the irrigation amount, and a mobile soil moisture recorder measures the soil moisture; each intelligent facility is connected to the Internet of Things system to realize digital monitoring and control, and to monitor the growth of paddy field crops in real time, providing a data acquisition and irrigation system for the present invention.

[0055] In the specific implementation process of the above embodiment, first, data on the dynamic changes of crop growth indicators, soil moisture conditions and meteorological factors are collected to obtain a joint data set of crop irrigation; by constructing a system perception layer, it gets rid of the reliance on a single soil moisture condition to determine irrigation needs, focuses on the dynamic correlation between meteorological changes and crop growth stages, and uses multi-source data fusion to lay the foundation for future efficient water-saving irrigation of paddy fields; secondly, the adaptive irrigation threshold is calculated based on the joint data set of crop irrigation, that is, irrigation is selected by comparing the crop drought tolerance critical value. When the soil moisture content is greater than the adaptive irrigation threshold, a hybrid prediction model is constructed to output the rainfall prediction value and the rainfall prediction time, and the irrigation amount prediction value and the irrigation time prediction value are calculated; this method uses the hybrid prediction model to indirectly obtain the irrigation prediction value, integrates the SVM classification model, the attention mechanism and the BiLSTM neural network, identifies the weights of various meteorological factors, greatly improves the stability and accuracy of the prediction, is suitable for the accurate prediction of extreme rainfall events, and greatly increases the adaptability in extreme weather. At the same time, the accurate prediction results are helpful for the subsequent calculation of the soil irrigation amount difference and the calculation of the irrigation time prediction value. Irrigation volume correction achieves better irrigation water-saving effects. When soil moisture content is less than or greater than the adaptive irrigation threshold, a phased control strategy is adopted for phased irrigation. That is, irrigation is carried out in different regions according to the different water requirements and rainfall at different crop growth stages. The phased strategy reduces irrigation volume while ensuring normal crop growth, thereby improving water conservation. The spatiotemporal precision irrigation method is then combined with the spatiotemporal precision irrigation method. By using a small amount of multiple irrigation strategy as a temporal precision irrigation method, a spatial precision irrigation method is obtained by considering the slope. This method reduces gravity water loss through spatiotemporal precision water distribution, significantly improves irrigation uniformity, and achieves efficient water-saving irrigation while ensuring yield increase. Finally, after precision irrigation, the irrigation volume is corrected considering soil irrigation volume differences, achieving more precise control of irrigation volume and realizing refined irrigation control management. The dry season-rainy season intelligent switching strategy is then used for dynamic water storage, and real-time warning and remote monitoring of water levels are then carried out. This method prioritizes high-priority areas and implements intelligent drainage-water storage linkage to achieve control in wet seasons and accurate irrigation in dry seasons, resulting in stable water-saving effects. The whole process forms a triple optimization mechanism of spatial dynamic partitioning + time stage cutting + seasonal strategy superposition, which has the value of large-scale industrial promotion.

[0056] Further, in order to better introduce the technical solution of the embodiment of the present invention, based on the above-mentioned intelligent and efficient water-saving irrigation system for paddy fields, as shown in FIG. Figure 2 As shown, the embodiment of the present invention provides an intelligent and efficient water-saving irrigation method for paddy fields, which specifically includes the following contents:

[0057] S1. Select crop planting areas and collect crop growth indicators, soil moisture and weather data.

[0058] The data of dynamic changes of factors are used to generate a joint data set of crop irrigation by fusion of multi-source data;

[0059] Said S1 comprises the following steps:

[0060] S11. Select a crop planting area and measure the leaf size of the crops in the crop planting area.

[0061] Length and width of leaves, calculate the leaf area of ​​crops ,in Indicates the leaf length of crops. Indicates the leaf width of crops; then count the number of crops per unit area in the crop planting area and calculate the leaf area index of crops ,in Indicates the number of crops per unit area;

[0062] A set of leaf area indices of crops is set, wherein the leaf area indices of the crops in the set correspond to the crop growth stages, the leaf area index of the crop corresponding to the leaf area index of the crop is found, the crop growth stage corresponding to the leaf area index of the crop is obtained, and each crop growth stage in the crop planting area is marked; the same and adjacent crop growth stages in the crop planting area are grouped into the same crop planting sub-area, to obtain a plurality of the same crop planting sub-areas, and the same crop growth stages are numbered;

[0063] The soil moisture sensor is used to measure the soil moisture in the same crop planting sub-area, and the soil moisture corresponding to several crop planting sub-areas is obtained to form a crop soil moisture set. ,in Indicates the Soil moisture corresponding to each crop planting sub-area; then obtaining meteorological data and time characteristics of the crop planting areas in previous years, wherein the meteorological data includes temperature, air pressure, wind speed, humidity and evaporation, and the time characteristics include season, month and week number, to form a meteorological data set and a time characteristic set of the crop planting areas in previous years;

[0064] S12: Record the label corresponding to the same crop planting sub-area to obtain a crop planting sub-area set. ,in Indicates the The crop growth stage corresponding to each crop planting sub-area is combined with the crop planting sub-area set, the crop soil moisture set and the crop planting area meteorological data set in previous years to form a crop irrigation joint data set. ,in represents the meteorological data set of crop planting areas in previous years, Represents the temporal characteristics of crop planting areas in previous years;

[0065] In this embodiment, a joint crop irrigation data set is obtained by collecting data on the dynamic changes of crop growth indicators, soil moisture conditions, and meteorological factors. By building a system perception layer, the system can get rid of the reliance on a single soil moisture condition to determine irrigation needs, focus on the dynamic correlation between meteorological changes and crop growth stages, and use multi-source data fusion to lay the foundation for future efficient water-saving irrigation of paddy fields. Specifically, for example, a standardized rice field is selected as the monitoring area, 10 rice plants are randomly selected, their leaf lengths and widths are measured, and the average value is taken to obtain a single leaf area of ​​49.8 cm 2 , the rice plant density per unit area (1 square meter) was calculated to be 200 plants, and the leaf area index was calculated to be 0.996; a rice leaf area index set was set, including the tillering stage [0.8, 1.2], the jointing stage [1.3, 2.0], and the heading stage [2.1, 3.0]. Corresponding to the tillering stage, continuous areas in the tillering stage were merged into three planting sub-areas and marked as "tillering stage-1", "tillering stage-2", and "tillering stage-3"; then, the soil moisture of crops was measured, and meteorological data and time characteristics of the past five years were selected (marked as summer / July / third week, etc.) to form a joint crop irrigation data set to provide multi-dimensional data support;

[0066] S2. Calculating an adaptive irrigation threshold based on the crop irrigation demand data, comparing the soil moisture content with the adaptive irrigation threshold, and predicting the irrigation time and amount based on a neural network when the soil moisture content is greater than the adaptive irrigation threshold to obtain an irrigation prediction value;

[0067] The S2 comprises the following steps:

[0068] S21. Determine the water content of crops when they wilt and the maximum water holding capacity of crops in the crop planting sub-area, and set a dynamic weight coefficient set for the crop growth stage corresponding to the crop planting sub-area. ,in Indicates the The dynamic weight coefficient corresponding to the growth stage of the crop; the adaptive irrigation threshold corresponding to the growth stage of the nth crop is calculated ,in Indicates the water content of crops in the crop planting sub-area when they wilt. It represents the maximum water holding capacity of crops in the crop planting sub-area;

[0069] S22. Calculate the soil moisture content in the crop planting sub-area set based on the crop soil moisture set in the crop irrigation joint data set. For different crop planting sub-areas, compare the soil moisture content with the adaptive irrigation threshold. When the soil moisture content is greater than the adaptive irrigation threshold, construct an SVM-Attention-BiLSTM hybrid prediction model to predict rainfall time and rainfall amount, and output a rainfall amount prediction value and rainfall prediction time. The specific steps are as follows:

[0070] S221, based on the meteorological data set of the crop planting area in previous years and the time feature set of the crop planting area in previous years in the crop irrigation joint data set, the meteorological data set of the crop planting area in previous years is normalized, and a sliding window is set, and the meteorological data of the crop planting area in previous years in the sliding window is used to generate rainfall samples to form a rainfall sample set; the time feature set of the crop planting area in previous years is normalized according to the step size Divide into rainfall time sets;

[0071] S222. Set the BiLSTM to include an input layer, a forward LSTM, a backward LSTM, and an output layer. After adding the Attention mechanism to the backward LSTM, and adding the pooling layer to the Attention mechanism, add the SVM classification model to the pooling layer, and the output layer after the pooling layer to construct a hybrid prediction model; divide the rainfall sample set into a sample training set and a sample test set, and divide the rainfall time set to obtain a sample training time set and a sample test time set, input the sample training set and the sample training time set into the hybrid prediction model, and the Attention mechanism performs feature learning on the temperature, air pressure, wind speed, humidity, and evaporation, generates time step weights and aggregates them, and the SVM classification model outputs whether it will rain in the next 30 days, using 1 to indicate rainfall and 0 to indicate no rainfall. When it rains, the output layer outputs the rainfall prediction result, and when it does not rain, the output layer outputs 0; until the hybrid prediction model converges, a trained hybrid prediction model is obtained;

[0072] S223. The sample test set and the sample test time set are input into the trained hybrid prediction model. An accuracy threshold is set. When the accuracy of the output rainfall prediction result is greater than the accuracy threshold, the iteration is stopped to obtain the SVM-Attention-BiLSTM hybrid prediction model. Otherwise, the weight is adjusted until the accuracy of the rainfall prediction result is greater than the accuracy threshold.

[0073] S23. Before getting the current time The temperature, air pressure, wind speed, humidity, evaporation and the corresponding season, month and week numbers of the month are used to obtain the current meteorological data set and the current time feature set. The current meteorological data set is normalized and the current time feature set is subjected to sliding window processing, and then input into the SVM-Attention-BiLSTM hybrid prediction model to output the rainfall prediction value and rainfall prediction time;

[0074] The irrigation amount at the current time is obtained again, and the rainfall prediction value within the range from the current time to the rainfall prediction time is counted to obtain the irrigation amount prediction value and the irrigation time prediction value, and the irrigation amount prediction value and the irrigation time prediction value are recorded as the irrigation prediction value;

[0075] In this embodiment, an adaptive irrigation threshold is calculated based on a joint crop irrigation data set, that is, irrigation is selected by comparing the crop drought tolerance threshold. When the soil moisture content is greater than the adaptive irrigation threshold, a hybrid prediction model is constructed to output a rainfall prediction value and a rainfall prediction time. The hybrid prediction model is used to indirectly obtain the irrigation prediction value. The SVM classification model, the attention mechanism, and the BiLSTM neural network are integrated to identify the weights of various meteorological factors, greatly improving the stability and accuracy of the prediction. It is suitable for the accurate prediction of extreme rainfall events and greatly increases the adaptability in extreme weather. At the same time, the accurate prediction results are helpful for the subsequent calculation of the soil irrigation amount difference and the irrigation amount correction, thereby achieving a better irrigation water-saving effect. Specifically, for example, the soil moisture content of rice is 25% when it wilts, and the maximum water holding capacity of the rice field is 45%. Rice is divided into five growth stages, and dynamic weights are set: 0.3 for the seedling stage, 0.6 for the tillering stage, 0.8 for the jointing and booting stage, 0.5 for the heading and flowering stage, and 0.2 for the maturity stage. At this time, the tillering stage threshold is 37%; input data : Meteorological data (temperature, air pressure, wind speed, humidity, evaporation) and time characteristics (season, month, week number) for the past five years, generating rainfall samples in a 30-day window, dividing rainfall time sets into 7-day time steps, BiLSTM layer: bidirectional learning of meteorological data time series characteristics, Attention mechanism: focusing on key variables (such as humidity and evaporation), generating time step weights, SVM classifier: judging whether there will be rainfall in the next 30 days based on aggregated features; then input data: meteorological data of the previous three months (normalized) and time characteristics For the query (Summer / July / Week 2), the output is: 1 (Rainfall in the next 30 days), 20 mm of rain on the 10th day, and 100 mm of rain on the 25th day; the average daily water consumption during the tillering period is 5 mm, and the irrigation required in the next 10 days without rain is 50 mm. There is 20 mm of rain on the 10th day, and the predicted irrigation amount is 30 mm. The actual rainfall on the 10th day is 19 mm, and the actual rainfall on the 25th day is 105 mm. The prediction is accurate for the rainfall on the 10th day and highly adaptable to the heavy rain on the 25th day under extreme weather conditions.

[0076] S3. Compare the irrigation prediction value and the current irrigation value to obtain the difference in soil irrigation amount. When the soil moisture content is less than or greater than the adaptive irrigation threshold, adopt a staged control strategy for staged irrigation, and then combine the spatiotemporal precision irrigation method to accurately irrigate crops to achieve efficient water-saving irrigation;

[0077] The S3 includes the following steps:

[0078] S31, the irrigation prediction value includes several irrigation time prediction values ​​and corresponding irrigation amount prediction values, finding the rated irrigation amount from the current time to the irrigation time prediction value, calculating the difference between the rated irrigation amount and the irrigation amount prediction value, and recording it as the soil irrigation amount difference;

[0079] S32. When the soil moisture content is less than or greater than the adaptive irrigation threshold, a staged control strategy is used for staged irrigation in different crop planting sub-areas. The specific steps are as follows:

[0080] S321. Determine, based on the crop planting sub-region set in the crop irrigation joint data set, the crop growth stages corresponding to the crop planting sub-region set, set the crop growth stages to be transplanting stage, tillering stage, jointing stage, heading stage, and maturity stage to obtain a crop growth stage set, and then determine the water requirements of different crop growth stages to obtain a crop growth stage water requirement set;

[0081] S322. Arrange irrigation equipment in the crop planting area, the irrigation equipment including a main pipe, branch pipes, a flow meter, a water gauge and a water level sensor, etc.; determine, for different crop planting sub-areas in the crop planting area, the crop soil moisture of each of the different crop planting sub-areas based on the crop soil moisture set in the crop irrigation joint data set, and calculate the soil moisture content of each of the different crop planting sub-areas;

[0082] S323. Using a staged control strategy, when the crop growth stage of the crop planting sub-area is in the transplanting stage, the crop transplanting stage water requirement in the crop growth stage water requirement set is used to subtract the soil moisture content of the crop planting sub-area to obtain the crop transplanting stage irrigation amount. When the crop growth stage of the crop planting sub-area is in the tillering stage, the crop tillering stage water requirement in the crop growth stage water requirement set is used to subtract the soil moisture content of the crop planting sub-area to obtain the crop tillering stage irrigation amount. The irrigation amounts for all crop growth stages are calculated sequentially to obtain the irrigation amounts for different crop planting sub-areas.

[0083] At this time, use irrigation equipment for irrigation;

[0084] S33. During the irrigation process, a small amount of multiple irrigations strategy is used as a temporal precision irrigation method. The slope is then taken into account to obtain a spatial precision irrigation method. The temporal and spatial precision irrigation method is used to precisely irrigate crops. The specific steps are as follows:

[0085] S331, dividing the irrigation amounts of different crop planting sub-areas according to the maximum water holding capacity of the crops in the crop planting sub-areas to obtain the irrigation frequency;

[0086] Set an intermittent time, and during the irrigation process using the irrigation equipment, irrigate the crop planting sub-area at the maximum water holding capacity of the crop at each intermittent time, and record the water level of the crop planting sub-area;

[0087] S332: Calculate the slope of the crop planting sub-area based on the water level of the crop planting sub-area, set the slope level and the corresponding correction coefficient, and obtain the slope position of the crop planting sub-area, where the slope position includes a flat slope, a slope top, and a slope bottom;

[0088] Set the upper and lower water level thresholds. When the crop planting sub-area is on a flat slope, irrigation will be stopped until the irrigation times are reached or the water level in the crop planting sub-area is greater than the lower water level threshold and less than the upper water level threshold.

[0089] When the crop planting sub-area is at the top of the slope, the irrigation frequency is reduced or the lower and upper water level thresholds are reduced; when the crop planting sub-area is at the bottom of the slope, the irrigation frequency is increased or the lower and upper water level thresholds are increased to achieve precise irrigation of crops.

[0090] Specifically, the reduced number of irrigation times is equal to 1 minus the absolute value of the correction coefficient multiplied by the number of irrigation times, and the reduced lower limit of the water level threshold and the upper limit of the water level threshold are equal to 1 minus the absolute value of the correction coefficient multiplied by the lower limit of the water level threshold and the upper limit of the water level threshold; the increased number of irrigation times is equal to 1 minus the absolute value of the correction coefficient multiplied by the number of irrigation times, and the increased lower limit of the water level threshold and the upper limit of the water level threshold are equal to 1 minus the absolute value of the correction coefficient multiplied by the lower limit of the water level threshold and the upper limit of the water level threshold;

[0091] In this embodiment, when the soil moisture content is less than or greater than the adaptive irrigation threshold, a staged control strategy is adopted for staged irrigation, that is, for different crop growth stages, the water demand and rainfall are different, and irrigation is carried out separately in different regions. The staged strategy reduces the irrigation amount while ensuring the normal growth of crops, thereby improving the water saving rate; the spatiotemporal precision irrigation method is then combined with the precise irrigation of crops. By using a small amount of multiple irrigation strategy as a time precision irrigation method, the spatial precision irrigation method is obtained by considering the slope. This method reduces gravity water loss through precise spatiotemporal water distribution, significantly improves irrigation uniformity, and ensures increased yields. At the same time, efficient water-saving irrigation is achieved; specifically, for example, the rated irrigation volume is set to 40mm, at this time the soil irrigation volume difference is 10mm; the water requirement for each growth stage of rice is set: 120mm in transplanting period, 150mm in tillering period, 180mm in jointing period, 200mm in heading period, and 100mm in maturity period; set sub-area A (transplanting period), the measured soil moisture content is 100mm, the required irrigation volume is 120mm-100mm=20mm, sub-area B (heading period), the measured soil moisture content is 170mm, the required irrigation volume is 200mm-170mm=30mm, and staged irrigation is implemented. Calculated per mu, the original irrigation volume was 200 cubic meters, and the actual irrigation volume was 180 cubic meters, a 10% decrease in irrigation volume; Sub-area A: flat slope (slope correction coefficient 0), Sub-area B: top of slope (slope 2°, slope correction coefficient +0.2), Sub-area C: bottom of slope (slope 3°, slope correction coefficient -0.3); Sub-area B originally planned to irrigate twice, and after adjustment, the number remained unchanged, with the lower limit reduced from 5cm to 4cm, and the upper limit from 10cm to 8cm; Sub-area C originally planned to irrigate three times, and after adjustment, the number was twice, with the lower limit increased from 5cm to 6.5cm, and the upper limit increased from 10cm to 13cm. m; at this time, irrigation in the flat slope area (sub-area A) stops when the water level reaches a threshold (5-10 cm); irrigation in the top slope area (sub-area B) stops after the water level reaches 4-8 cm after two irrigations; in the bottom slope area (sub-area C), the single irrigation volume is increased to a range of 6.5-13 cm. Taking into account the slope change, the water flow path can be shortened, reducing water leakage and evaporation losses caused by long-term retention. Throughout the entire process, calculated per mu, the original irrigation volume was 200 cubic meters, and the actual irrigation volume was 160 cubic meters, increasing the overall irrigation efficiency by 20%, achieving a balance between rice growth water requirements and resource conservation, and achieving efficient water-saving irrigation.

[0092] S4. After precise irrigation, taking into account the differences in soil irrigation amounts, a dry season-rainy season intelligent switching strategy is used to dynamically store water, and real-time early warning and remote monitoring of water levels are performed;

[0093] The S4 comprises the following steps:

[0094] S41. After precision irrigation, the remaining water volume in the crop planting area is obtained, and a water volume threshold is set. When the remaining water volume is less than the water volume threshold, a dry season strategy is introduced. Priorities are assigned to the crop growth stages to obtain the crop growth stage priorities. The crop planting sub-areas are irrigated in descending order of their growth stage priorities, and the irrigation volume of the crop planting sub-areas is added to the soil irrigation volume difference. Return water is collected, sedimented, filtered, and disinfected before being re-injected into the reservoir.

[0095] When the remaining water volume is greater than or equal to the water volume threshold, the crop planting area enters the rainy season; the irrigation volume of the crop planting sub-area is subtracted from the soil irrigation volume difference, and the crop planting sub-areas are irrigated in turn; a time threshold and a water holding capacity threshold are set, and when the maximum water holding capacity of the crop is greater than the water holding capacity threshold and the time is maintained for more than the time threshold, the drainage pump is started to drain the water into the reservoir;

[0096] S42: Real-time monitoring of the water level in the crop planting sub-area. When the water level in the crop planting sub-area is less than or equal to a lower water level threshold or greater than or equal to an upper water level threshold, an automatic alarm is triggered, and the position of the crop planting sub-area is located.

[0097] In this embodiment, after precise irrigation, the irrigation amount is corrected by considering the difference in soil irrigation amount, so as to control the irrigation amount more accurately, avoid water resource waste, realize refined management of irrigation control, use the intelligent switching strategy of dry season-rainy season for dynamic water storage, and then carry out real-time warning and remote monitoring of water level. This method gives priority to meeting the needs of high-priority areas, and implements intelligent linkage of drainage and water storage at the same time, so as to achieve control in wet times and accurate irrigation in dry times, and achieve stable water-saving effect. Specifically, for example, a water volume threshold of 2000 cubic meters, a time threshold of 3 days (continuously exceeding water holding capacity), and a water holding capacity threshold of 90% are set; in the dry season (remaining water volume 1500m³<threshold value), the priority of rice growth stage (from high to low) is heading and flowering stage (sub-area A), jointing and booting stage (sub-area B), and tillering stage (sub-area C); differentiated irrigation, in order of priority, each sub-area is increased by 10mm (soil holding capacity) on the basic irrigation amount. The system collects irrigation return water, sediments it, filters it, and disinfects it with ultraviolet light before replenishing it to the reservoir. During the rainy season (remaining water volume ≥ a threshold of 2,500 m³), ​​differentiated irrigation is implemented, with each sub-area receiving a -10 mm increase over the base irrigation volume. When the soil water holding capacity in a sub-area is monitored to be >90% for three consecutive days, the drainage pump automatically activates. When the water level in sub-area A reaches 15 mm (<20 mm lower limit), a water shortage alarm is triggered, locating the coordinates of area A. When the water level in sub-area B reaches 85 mm (≥80 mm upper limit), a flood alarm is triggered, locating the coordinates of area B, and the system sends an alarm message to the management platform. Dynamic water recycling ensures stable water conservation. Throughout the entire process, rainy season water storage accounts for 15% of the total reservoir volume, avoiding water waste. Furthermore, the irrigation volume corrected for soil irrigation volume differences (55 mm - 10 mm) more closely matches the actual irrigation volume (45 mm), contributing to efficient water conservation.

[0098] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0099] The preferred embodiments of the invention disclosed above are intended only to help illustrate the invention. These preferred embodiments do not exhaust all details, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. Intelligent and efficient water-saving irrigation system for paddy fields, characterized by: include: Multi-source data fusion module is used to collect dynamic data of crop growth indicators, soil moisture and meteorological factors. The data of dynamic changes are used to generate a joint data set of crop irrigation by fusing multi-source data; The irrigation prediction module is used to calculate the adaptive irrigation threshold, compare the soil moisture content with the adaptive irrigation threshold, and predict the irrigation time and amount based on the neural network when the soil moisture content is greater than the adaptive irrigation threshold to obtain the irrigation prediction value; The efficient water-saving irrigation module is used to calculate the difference in soil irrigation amount based on the irrigation prediction value and the current irrigation value. When the soil moisture content is less than or greater than the adaptive irrigation threshold, a staged control strategy is adopted for staged irrigation, and then the crops are precisely irrigated in combination with the spatiotemporal precision irrigation method. An early warning and monitoring module is used to take into account the differences in soil irrigation volume after precise irrigation, use a dry season-rainy season intelligent switching strategy for dynamic water storage, and provide real-time early warning and remote monitoring of water levels; The precise irrigation of crops by combining the spatiotemporal precise irrigation method includes: During the irrigation process, a small amount of multiple irrigation strategy is used as a time-precision irrigation method. The irrigation amount of different crop planting sub-areas is divided according to the maximum water holding capacity of the crops in the crop planting sub-areas to obtain the irrigation frequency. Set an intermittent time, and during the irrigation process using the irrigation equipment, irrigate the crop planting sub-area at the maximum water holding capacity of the crop at each intermittent time, and record the water level of the crop planting sub-area; A spatial precision irrigation method is obtained by taking the slope into consideration, calculating the slope of the crop planting sub-area according to the water level of the crop planting sub-area, setting the slope level and the corresponding correction coefficient, and obtaining the slope position of the crop planting sub-area; The slope position includes a flat slope, a slope top and a slope bottom; Set the upper and lower water level thresholds. When the crop planting sub-area is on a flat slope, irrigation will be stopped until the irrigation times are reached or the water level in the crop planting sub-area is greater than the lower water level threshold and less than the upper water level threshold. When the crop planting sub-area is at the top of the slope, reduce the number of irrigations or reduce the lower and upper limits of the water level threshold; when the crop planting sub-area is at the bottom of the slope, increase the number of irrigations or increase the lower and upper limits of the water level threshold.

2. The intelligent and efficient water-saving irrigation system for paddy fields according to claim 1 is characterized in that: The data collected on the dynamic changes of crop growth indicators, soil moisture conditions and meteorological factors include: Selecting a crop planting area, measuring the leaf area of ​​the crops in the crop planting area, and calculating the leaf area index of the crops; Set a set of crop leaf area indices, find the crop leaf area index corresponding to the crop leaf area index, obtain the crop growth stage corresponding to the crop leaf area index, and group the same and adjacent crop growth stages in the crop planting area into the same crop planting sub-area, thereby obtaining several identical crop planting sub-areas; The soil moisture in the same crop planting sub-area is measured to obtain a crop soil moisture set; then the meteorological data and time characteristics of the crop planting area in previous years are obtained to form a meteorological data set of the crop planting area in previous years and a time characteristic set of the crop planting area in previous years.

3. The intelligent and efficient water-saving irrigation system for paddy fields according to claim 2 is characterized in that: The method of generating a crop irrigation joint data set by fusing multi-source data includes: The same crop planting sub-region is recorded as a crop planting sub-region set, and is combined with the crop soil moisture set and the crop planting region meteorological data set of previous years to form a crop irrigation joint data set.

4. The intelligent and efficient water-saving irrigation system for paddy fields according to claim 3 is characterized in that: The calculating of the adaptive irrigation threshold comprises: The water content of crops when they wilt and the maximum water holding capacity of crops in the crop planting sub-area are measured. According to the crop growth stage corresponding to the crop planting sub-area, a dynamic weight coefficient is set, and then the adaptive irrigation threshold corresponding to the crop growth stage is calculated.

5. The intelligent and efficient water-saving irrigation system for paddy fields according to claim 4 is characterized in that: When the soil moisture content is greater than the adaptive irrigation threshold, predicting the irrigation time and amount based on the neural network includes: According to the crop irrigation joint data set, the soil moisture content in the crop planting sub-area set is calculated. For different crop planting sub-areas, the soil moisture content and the adaptive irrigation threshold are compared. When the soil moisture content is greater than the adaptive irrigation threshold, the meteorological data set of the crop planting area in previous years is normalized, and a sliding window is set. The meteorological data of the crop planting area in previous years in the sliding window are used to generate rainfall samples to form a rainfall sample set; the time feature set of the crop planting area in previous years is normalized according to the step size. Divide into rainfall time sets; The BiLSTM is set to include an input layer, a forward LSTM, a backward LSTM, and an output layer. After the Attention mechanism is added to the backward LSTM, and the pooling layer is added to the Attention mechanism, the SVM classification model is added to the pooling layer, and the output layer is after the pooling layer to construct a hybrid prediction model. The rainfall sample set is divided into a sample training set and a sample test set, and the rainfall time set is divided to obtain a sample training time set and a sample test time set. The sample training set and the sample training time set are input into the hybrid prediction model. The Attention mechanism performs feature learning on the temperature, air pressure, wind speed, humidity, and evaporation, generates time step weights, and aggregates them. The SVM classification model outputs whether it will rain in the next 30 days, using 1 to indicate rainfall and 0 to indicate no rainfall. When rainfall occurs, the output layer outputs the rainfall prediction result, and when no rainfall occurs, the output layer outputs 0. This is continued until the hybrid prediction model converges to obtain a trained hybrid prediction model. Then, the sample test set and the sample test time set are input into the trained hybrid prediction model. The accuracy threshold is set. When the accuracy of the output rainfall prediction result is greater than the accuracy threshold, the iteration is stopped to obtain the SVM-Attention-BiLSTM hybrid prediction model. Otherwise, the weight is adjusted until the accuracy of the rainfall prediction result is greater than the accuracy threshold. Get the current time The temperature, air pressure, wind speed, humidity, evaporation and the corresponding season, month and week numbers of the month are used to obtain the current meteorological data set and the current time feature set. The current meteorological data set is normalized and the current time feature set is subjected to sliding window processing, and then input into the SVM-Attention-BiLSTM hybrid prediction model to output the rainfall prediction value and rainfall prediction time; The irrigation amount at the current time is obtained again, and the rainfall forecast value within the range from the current time to the rainfall forecast time is counted, and the irrigation amount forecast value and the irrigation time forecast value are calculated and recorded as the irrigation forecast value.

6. The intelligent and efficient water-saving irrigation system for paddy fields according to claim 5, characterized in that: The calculation of the soil irrigation amount difference, when the soil moisture content is less than or greater than the adaptive irrigation threshold, adopting a staged control strategy for staged irrigation includes: The irrigation prediction value includes a number of irrigation time prediction values ​​and corresponding irrigation amount prediction values, finding the rated irrigation amount from the current time to the irrigation time prediction value, calculating the difference between the rated irrigation amount and the irrigation amount prediction value, and recording it as the soil irrigation amount difference; When the soil moisture content is less than or greater than the adaptive irrigation threshold, determining the crop growth stage corresponding to the set of crop planting sub-areas for different crop planting sub-areas, and then calculating the water requirements of the different crop growth stages; calculating the soil moisture content of the different crop planting sub-areas based on the crop irrigation joint data set; The irrigation amount at different crop growth stages is calculated in sequence to obtain the irrigation amount for different crop planting sub-areas, and irrigation equipment is used for irrigation.

7. The intelligent and efficient water-saving irrigation system for paddy fields according to claim 1, characterized in that: The use of the dry season-rainy season intelligent switching strategy for dynamic water storage includes: After precision irrigation, the remaining water volume in the crop planting area is obtained and a water volume threshold is set. When the remaining water volume is less than the water volume threshold, the dry season strategy is introduced. The crop growth stage priority is set. At this time, the crop planting sub-areas are irrigated in descending order according to the crop growth stage priority, and the irrigation volume of the crop planting sub-areas is added to the soil irrigation volume difference. The return water is collected, sedimented, filtered and disinfected, and then re-injected into the reservoir. When the remaining water volume is greater than or equal to the water volume threshold, the rainy season strategy is introduced; the irrigation volume of the crop planting sub-area is subtracted from the soil irrigation volume difference, and the crop planting sub-areas are irrigated in turn; a time threshold and a water holding capacity threshold are set, and when the maximum water holding capacity of the crop is greater than the water holding capacity threshold and the time is maintained for more than the time threshold, the drainage pump is started to drain the water into the reservoir.

8. The intelligent and efficient water-saving irrigation system for paddy fields according to claim 7, characterized in that: The real-time early warning and remote monitoring of water level include: The water level of the crop planting sub-area is monitored in real time. When the water level of the crop planting sub-area is not greater than the lower limit of the water level threshold or not less than the upper limit of the water level threshold, an automatic alarm is issued and the position of the crop planting sub-area is located.

9. An intelligent and efficient water-saving irrigation method for paddy fields, characterized in that: The intelligent and efficient water-saving irrigation system for paddy fields according to any one of claims 1 to 8 specifically comprises: S1. Select crop planting areas and collect data on crop growth indicators, soil moisture and meteorological factors. Dynamically changing data, using multi-source data fusion to generate a joint crop irrigation data set; S2. Calculating an adaptive irrigation threshold based on the crop irrigation demand data, comparing the soil moisture content with the adaptive irrigation threshold, and predicting the irrigation time and amount based on a neural network when the soil moisture content is greater than the adaptive irrigation threshold to obtain an irrigation prediction value; S3. Compare the irrigation prediction value and the current irrigation value to obtain the difference in soil irrigation amount. When the soil moisture content is less than or greater than the adaptive irrigation threshold, adopt a staged control strategy for staged irrigation, and then combine the spatiotemporal precision irrigation method to accurately irrigate crops to achieve efficient water-saving irrigation; S4. After precise irrigation, considering the difference in soil irrigation amount, a dry season-rainy season intelligent switching strategy is used to dynamically store water, and real-time early warning and remote monitoring of the water level are carried out.

Images