Intelligent irrigation system for orchard in arid area

Abstract

The invention is suitable for the technical field of intelligent irrigation, and provides an intelligent irrigation system for an orchard in an arid region, which comprises an orchard image data acquisition end, a crop growth data acquisition end, a cloud platform, a remote control end and an irrigation device, and is characterized in that the cloud platform is connected with the orchard image data acquisition end, the crop growth data acquisition end and the remote control end; the remote control end is connected with the irrigation device; the cloud platform comprises an arid region monitoring module and a crop water demand prediction module; after the cloud platform receives data transmitted by the orchard image data acquisition end and the crop growth data acquisition end, the data is analyzed through the arid area monitoring module and the crop water demand prediction module, and then an analysis result is sent to the remote control end, so that the remote control end obtains a corresponding irrigation strategy after analysis and processing. And the irrigation device is controlled to perform intelligent irrigation based on the irrigation strategy, so that the problems of water resource waste and large energy consumption in intelligent irrigation are solved.

Description

Technical Field

[0001] The invention relates to the technical field of intelligent irrigation, and in particular to an intelligent irrigation system for orchards in arid areas. Background Art

[0002] Smart irrigation is an advanced agricultural irrigation method. It has the advantages of automation, precision, and intelligence. It can improve agricultural production efficiency, save water and environmental resources, and improve farmers' economic benefits and the yield and quality of crops. Proper irrigation has many benefits for fruit trees, while unreasonable irrigation will have a negative impact on the growth and results of fruit trees.

[0003] In the prior art, most of the intelligent irrigation of orchards is based on soil weather conditions, such as soil temperature, air humidity and air temperature, and other factors that affect demand are not analyzed based on multiple dimensions. Chinese invention patent CN118696809A discloses a method for fine irrigation of hilly orchards, including monitoring to obtain a low-resolution soil moisture distribution map; drones collect orchard DOM and DSM; identify ground points from DOM and correspond to DSM and interpolate to generate DEM; extract single fruit tree codes from DOM and generate fruit tree area vector files; obtain soil monitoring result maps around single fruit trees with DEM, low-resolution soil moisture distribution maps, and fruit tree area vector files; correspond fruit tree codes to solenoid valves of irrigation devices; generate fine irrigation prescription maps with drought monitoring result maps to control solenoid valve irrigation. It can be seen that the patent does not disclose the differences in the amount of water required for each crop or plant, and does not take into account the differences caused by the uncertainty of weather factors affecting the soil, so that the final irrigation strategy is not accurate enough, and there may be problems such as some water resources not being fully utilized and high energy consumption. Summary of the invention

[0004] The embodiment of the present application provides an intelligent irrigation system for orchards in arid areas, which is used to solve the problems of water resource waste and high energy consumption in intelligent irrigation.

[0005] A first aspect of an embodiment of the present application provides an intelligent irrigation system for an orchard in an arid area, comprising:

[0006] An orchard image data collection terminal, a crop growth data collection terminal, a cloud platform, a remote control terminal and an irrigation device, wherein the cloud platform is respectively connected to the orchard image data collection terminal, the crop growth data collection terminal and the remote control terminal, and the remote control terminal is connected to the irrigation device;

[0007] The orchard image data acquisition terminal uses a drone to take real-time image information of crops in the orchard, and uploads the captured images to the cloud platform;

[0008] The crop growth data collection terminal collects multi-dimensional growth data of crops through sensors, and uploads the collected data to the cloud platform;

[0009] The cloud platform includes a drought area monitoring module and a crop water requirement prediction module. The drought area monitoring module is used to perform image recognition on image information of crops in an orchard to determine potential drought areas in the orchard; the crop water requirement prediction module is used to analyze and predict growth data of crops in drought areas to obtain water requirements of crops in drought areas within a preset time; the cloud platform transmits the analysis results generated by the drought area monitoring module and the crop water requirement prediction module to the remote control terminal;

[0010] The remote control terminal is used to analyze the analysis results transmitted by the cloud platform, determine the irrigation strategy corresponding to the drought area, and control the irrigation device to perform intelligent irrigation through the determined irrigation strategy.

[0011] Furthermore, the working method of the drought area monitoring module includes:

[0012] Grid division is performed based on the received image information, and data corresponding to each divided grid is preprocessed;

[0013] The preprocessed data is input into a preset orchard convolutional neural network, and a probability vector of each grid is output, wherein the probability vector includes drought elements and non-drought elements;

[0014] The distribution of all potential drought areas in the orchard is determined according to the probability vector.

[0015] Furthermore, the training process of the preset orchard convolutional neural network includes the following:

[0016] Collect historical orchard images taken by drones, which include information on the drought and non-drought states of orchard conditions in different seasons, at different times and under different weather conditions;

[0017] The collected historical orchard images are preprocessed and the preprocessed data is input into the constructed model, which includes the first convolutional layer, the second convolutional layer, the pooling layer, the first fully connected layer and the second fully connected layer;

[0018] The binary cross entropy loss function and optimizer are used for multiple iterative training. When the preset number of training rounds is reached, the model training is completed and the output is a probability vector containing arid areas and non-drought areas.

[0019] Furthermore, the working method of the crop water requirement prediction module includes:

[0020] Collect data on crop varieties, soil texture and meteorological data in arid areas;

[0021] A crop growth model is constructed based on the collected data, and the water requirement of crops in arid areas within a preset time is determined through the constructed crop growth model.

[0022] Furthermore, the crop growth model is constructed based on the collected data, and the water requirement of crops in arid areas within a preset time is determined by the constructed crop growth model, including:

[0023] Construct photosynthesis simulation model based on crop variety data and meteorological data in arid areas;

[0024] Construct a respiration simulation model based on crop variety data and meteorological data in arid areas;

[0025] Construct a transpiration simulation model based on crop variety data, soil texture data and meteorological data in arid areas;

[0026] A crop growth model is determined according to the constructed photosynthesis simulation model, respiration simulation model and transpiration simulation model, and the water requirement of crops in arid areas at different growth stages within a preset time is determined based on the crop growth model.

[0027] Furthermore, the expressions of the photosynthesis simulation model, the respiration simulation model and the transpiration simulation model include:

[0028] Photosynthesis simulation model:

[0029]

[0030] Where: P n is the net photosynthetic rate, P max is the maximum photosynthetic rate, I is the light intensity, K is the light intensity corresponding to the light saturation point, T is the actual temperature, T min 、T max and T opt They are the minimum temperature, maximum temperature and optimum temperature for photosynthesis respectively;

[0031] Respiration simulation model:

[0032]

[0033] Where: R is the respiratory rate, R m is the basal respiration rate at the reference temperature, T ref is the reference temperature;

[0034] Transpiration simulation model:

[0035]

[0036] Where: g s is the stomatal conductance, g s0 is the initial stomatal conductance, m is the sensitivity coefficient of stomatal to changes in photosynthetic rate and intercellular carbon dioxide concentration, A is the photosynthetic rate, C s is the intercellular carbon dioxide concentration.

[0037] Furthermore, the remote control terminal is used to analyze the analysis results transmitted by the cloud platform, determine the irrigation strategy corresponding to the drought area, and control the irrigation device to perform intelligent irrigation through the determined irrigation strategy, including:

[0038] The remote control terminal determines the precipitation within a preset time according to the collected weather environment information of the area where the orchard is located;

[0039] Calculate the target water requirement of the crops in the arid area within a preset time based on the determined precipitation and the water requirement of the crops in the arid area within a preset time;

[0040] The objective function is constructed with the shortest irrigation time and the minimum irrigation amount as the target parameters;

[0041] Determining constraints, wherein the constraints include a target water demand constraint for crops in arid areas, an irrigation unit performance constraint, a simultaneous operation quantity constraint for irrigation units, a water source water supply capacity constraint, and a soil water holding capacity constraint;

[0042] The irrigation strategy corresponding to each arid area is determined according to the objective function and the constraint conditions.

[0043] Furthermore, the calculation of the target water requirement of the crops in the arid area within the preset time based on the determined precipitation and the water requirement of the crops in the arid area within the preset time includes:

[0044] Calculate and determine the difference between precipitation and the water requirement of crops in arid areas within a preset time;

[0045] When the calculated difference meets the conditions, the target amount of water to be supplemented by the irrigation device is determined based on the water holding capacity of the soil in the arid area and the irrigation efficiency of the irrigation device.

[0046] Furthermore, the objective function is constructed with the shortest irrigation time and the smallest irrigation amount as the objective parameters, including:

[0047]

[0048] Where: n is the irrigation unit, T is the total irrigation time of the irrigation device, t i is the irrigation time of the i-th irrigation unit, Q is the total irrigation amount of the irrigation device, q iis the irrigation amount of the ith irrigation unit, Z is the total objective function value, w1 and w2 are the weight coefficients of T and Q respectively, and w1+w2=1.

[0049] Furthermore, the constraints include the target water demand constraint for crops in arid areas, the irrigation unit performance constraint, the number of irrigation units working simultaneously constraint, the water supply capacity constraint of the water source and the soil water holding capacity constraint, including:

[0050] Meeting target water demand constraints for crops in arid areas:

[0051]

[0052] Where: Q need The amount of water required for crops in dry areas within a preset time;

[0053] Irrigation unit performance constraints:

[0054] q i,min ≤q i ≤q i,max

[0055] t i,min ≤t i ≤t i,max

[0056] Where: q i,min and q i,max are the minimum and maximum allowable irrigation amounts for the ith irrigation unit, t i,min and t i，max are the shortest and longest irrigation time of the ith irrigation unit, respectively;

[0057] Constraints on the number of irrigation units working simultaneously:

[0058]

[0059] Where: m is the maximum number of irrigation units allowed to work simultaneously, x i is a 0-1 variable, x i =1 means the i-th irrigation unit is working, x i =0 means not working;

[0060] Constraints on water supply capacity:

[0061]

[0062] Where: Q source The maximum water supply per hour for the water source;

[0063] Soil water holding capacity constraints:

[0064]

[0065] Where: Q initiαl is the initial soil moisture content, Q soil The maximum water holding capacity of the soil.

[0066] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0067] The present invention uses a drone at an orchard image data acquisition end to shoot orchard crop images, and a sensor at a crop growth data acquisition end to collect crop growth data, and uploads the data to a cloud platform; the arid zone monitoring module and the crop water requirement prediction module in the cloud platform analyze and process the image and growth data respectively, obtain information on potential arid areas and crop water requirements in the orchard, and transmit the results to a remote control end; the remote control end determines the irrigation strategy for the arid area based on these analysis results, and then controls the irrigation device to implement intelligent irrigation. The entire process forms a closed-loop system from data acquisition, analysis to decision-making execution to adapt to the complex and changeable environment and crop growth requirements of orchards in arid areas. The intelligent irrigation system uses a variety of data acquisition methods, combined with the data analysis and processing capabilities of the cloud platform, and the intelligent decision-making of the remote control end, to achieve precise irrigation of orchards in arid areas, improve water resource utilization efficiency, and ensure the good growth of orchard crops. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 This is a structural diagram of an embodiment of an intelligent irrigation system for orchards in arid areas of the present invention;

[0069] Figure 2 This is a schematic diagram of an embodiment of the working method of the drought area monitoring module in the present invention;

[0070] Figure 3 This is a schematic flow chart of an embodiment of a working method of a crop water requirement prediction module in the present invention;

[0071] Figure 4 The figure is a flowchart of an embodiment of the working method of the remote control terminal in the present invention. DETAILED DESCRIPTION

[0072] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "corresponding to" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0073] Embodiment 1

[0074] The following is an introduction to the intelligent system for orchards in arid areas in this application from the perspective of data interaction between orchard image data collection, crop growth data collection, cloud platform, remote control, and irrigation devices. Figure 1 , an intelligent irrigation system for orchards in arid areas comprises:

[0075] Orchard image data collection terminal 1, crop growth data collection terminal 2, cloud platform 3, remote control terminal 4 and irrigation device 5, cloud platform 3 is respectively connected to orchard image data collection terminal 1, crop growth data collection terminal 2 and remote control terminal 4, remote control terminal 4 is connected to irrigation device 5;

[0076] The orchard image data collection terminal 1 uses a drone to take real-time images of crops in the orchard and uploads the captured images to a cloud platform 3;

[0077] The crop growth data collection terminal 2 collects multi-dimensional growth data of crops through sensors and uploads the collected data to the cloud platform 3;

[0078] The cloud platform 3 includes a drought area monitoring module 301 and a crop water requirement prediction module 302. The drought area monitoring module 301 is used to perform image recognition on image information of crops in the orchard to determine potential drought areas in the orchard; the crop water requirement prediction module 302 is used to analyze and predict the growth data of crops in the drought area to obtain the water requirement of crops in the drought area within a preset time; the cloud platform 3 transmits the analysis results generated by the drought area monitoring module and the crop water requirement prediction module to the remote control terminal 4;

[0079] The remote control terminal 4 is used to analyze the analysis results transmitted by the cloud platform 3, determine the irrigation strategy corresponding to the drought area, and control the irrigation device 5 to perform intelligent irrigation through the determined irrigation strategy.

[0080] Specifically, drones equipped with high-definition cameras can quickly and comprehensively obtain image information of orchard crops, with a wide coverage range and the ability to flexibly adjust shooting angles and areas, providing an intuitive data basis for subsequent drought area identification. A variety of sensors are used to collect multi-dimensional growth data of crops, including but not limited to soil moisture sensors, temperature sensors, light sensors, plant stem flow sensors, etc., to comprehensively reflect the growth status of crops from aspects such as soil environment, meteorological conditions, and the physiological indicators of crops themselves. The drought area monitoring module realizes automatic identification and positioning of potential drought areas, and can quickly and accurately determine the distribution and range of drought areas in orchards, providing target area information for subsequent precision irrigation. The crop water demand prediction module takes into account the impact of different crop varieties, growth stages, soil conditions, and meteorological factors on water demand. The model can dynamically adjust parameters and accurately predict the water demand of crops in drought areas within a preset time, providing key data support for irrigation decisions. After the remote control end receives the drought area and crop water requirement information transmitted by the cloud platform, it uses the optimization algorithm to build an irrigation strategy. The irrigation device has the ability to execute automatically and can automatically adjust the valve opening, water pump speed, etc. according to the received control instructions to achieve precise control of irrigation water volume and irrigation range. It also has a feedback function to transmit the operating status information of the irrigation device back to the remote control end for real-time monitoring and adjustment of the irrigation process.

[0081] Embodiment 2

[0082] See also Figure 2 , the working method of the drought area monitoring module includes the following steps:

[0083] S21. Gridding is performed based on the received image information, and the data corresponding to each grid is preprocessed;

[0084] Specifically, the orchard image captured by the drone is divided into multiple small grid areas. The grid size can be determined according to the size of the orchard, the image resolution, and the requirements for the positioning accuracy of the drought area. For example, the image is divided into square grids with a side length of 10-20 meters, and then each grid area is used as the input of the model, and the drought and non-drought probability vectors corresponding to each grid area are obtained in turn through the trained convolutional neural network.

[0085] S22. Input the preprocessed data into a preset orchard convolutional neural network, and output a probability vector for each grid, the probability vector including drought elements and non-drought elements;

[0086] The orchard convolutional neural network preset here refers to the trained convolutional neural network. The specific training process is as follows:

[0087] 1. Collect historical orchard images taken by drones, which include information on the drought and non-drought states of orchards in different seasons, at different times, and under different weather conditions;

[0088] Collect orchard images taken by drones. The data covers orchard conditions in different seasons, different times, and different weather conditions, including images of normal and drought conditions. At the same time, these images are manually annotated, and drought areas and non-drought areas are marked with labels to form training data sets, validation data sets, and test data sets. It should be noted that here we focus more on collecting image feature data related to drought, such as soil moisture, overall crop morphological changes, such as leaf withering and curling, changes in crown density, etc., as well as soil exposure and other drought-related features.

[0089] 2. Perform data preprocessing on the collected historical orchard images, and input the preprocessed data into the constructed model, which includes the first convolutional layer, the second convolutional layer, the pooling layer, the first fully connected layer, and the second fully connected layer;

[0090] Preprocessing includes image cropping and scaling, normalization, and data augmentation. The images are uniformly cropped or scaled to a fixed size to meet the input requirements of the convolutional neural network; the image pixel values ​​are normalized to the range of [0, 1] or [-1, 1]; and data augmentation operations are performed to increase data diversity and reduce the risk of overfitting.

[0091] The first convolution layer uses 32 convolution kernels of size 3×3, with a stride of 1 and padding of 1. The convolution kernel slides on the image, and convolution operations are performed on each 3×3 local area with the convolution kernel. The 32 convolution kernels here are used to extract low-level features of the image, such as edges, lines, and some simple texture information, focusing on basic features related to drought, such as the boundary between crops and soil in the orchard, the general outline of leaves, etc. The second convolution layer uses 64 3×3 convolution kernels with the same stride and padding as the first layer. This layer will further combine the features extracted in the first layer. Since there are more convolution kernels, more complex feature combinations can be learned, such as texture changes after leaves wither, cracked texture of soil caused by drought, and other feature patterns. These features are further abstracted based on the basic features of the first layer, and the extracted features focus more on the overall morphology and texture changes caused by drought.

[0092] After each convolutional layer, a 2×2 maximum pooling layer is connected with a stride of 2. The maximum pooling operation selects the maximum value in each 2×2 area as the output. In drought area identification, the pooling operation is to simplify the data and improve the computational efficiency while retaining the key features of drought. It is similar to the role of the pooling layer in other tasks, but it is aimed at retaining and extracting drought features.

[0093] After convolution and pooling operations, the feature map is flattened into a one-dimensional vector and then connected to two fully connected layers. The first fully connected layer has 256 neurons and uses the ReLU activation function. This layer will integrate the features and further extract the features. Through this layer, the features learned by the convolution layer are mapped to a higher-dimensional space to better classify drought and non-drought areas. This weight parameter is learned during the training process of orchard drought area identification and is optimized for drought characteristics. The second fully connected layer contains 2 neurons and uses the softmax activation function. This layer outputs the predicted probability of each category to determine whether the area in the image is a drought area.

[0094] 3. Use the binary cross entropy loss function and optimizer to perform multiple iterative training. When the preset number of training rounds is reached, the model training is completed and the output is a probability vector containing arid areas and non-drought areas.

[0095] Binary cross entropy is selected as the loss function to measure the difference between the model's predicted probability and the true label. The model is trained by minimizing the loss function so that the model's predicted results are as close to the actual situation as possible. The optimizer can automatically adjust the learning rate during the training process, so that the model converges faster and more stably. The preprocessed image data is divided into a training set, a validation set, and a test set. Multiple iterations of training are performed on the training set. In each iteration, the model performs forward propagation to calculate the predicted results based on the input image data, and then calculates the error between the predicted results and the true label through the loss function. The optimizer is then used to perform back propagation to update the model's parameters to reduce the error. When the preset number of training rounds is reached, the model training is completed, and the output is a probability vector containing arid areas and non-drought areas.

[0096] S23. Determine the distribution of all potential drought areas in the orchard based on the probability vector.

[0097] A threshold is set for each probability vector to determine whether the area is a drought area. For example, when the probability of a drought area output by the model is greater than 0.5, the grid area is determined to be a drought area. In this way, all areas are classified and marked to indicate which areas are drought areas and which are non-drought areas. According to the above classification results, mark them on the map of the orchard. Different colors are used here to represent drought areas and non-drought areas. For example, red is used to represent drought areas and green is used to represent non-drought areas. In this way, you can intuitively see the distribution of drought areas in the entire orchard. If more detailed information is needed, you can also mark the drought probability of each drought area on the map to indicate the difference in the degree of drought.

[0098] The above embodiment monitors the drought areas in the orchard and uses a convolutional neural network to identify the drought areas in the orchard based on the characteristics of the drought characteristics of the orchard, thereby effectively improving the accuracy of drought area monitoring.

[0099] Embodiment 3

[0100] See also Figure 3 The working method of the crop water requirement prediction module in the present invention includes the following:

[0101] S31. Collect data on crop varieties, soil texture and meteorological data in arid areas;

[0102] Specifically, after determining the drought area in the orchard in the above-mentioned embodiment 2, the crop water requirement prediction module here only predicts the water requirement for the crops in the identified drought area. The variety data of crops in the orchard reflects that different varieties of crops have different growth characteristics. For example, for the apple variety "Red Fuji", it is necessary to input its growth cycle, maximum leaf area index, duration of each growth stage, and root growth parameters. The parameters are used to describe the basic physiological and morphological characteristics of crops under ideal growth conditions. The soil texture data of the orchard is used to clarify the soil texture of the orchard, such as loam, sandy soil or clay. For loam, its water holding characteristics, aeration, etc. are different from other textures. Input the initial soil moisture content, consider the soil fertility factor, input the organic matter content, main nutrient content, etc. in the soil, and the data that affects the root development and overall growth status of the crop. The time interval of the meteorological data at the location of the orchard can be daily or hourly, for example, using hourly meteorological data, including hourly temperature, humidity, wind speed, solar radiation, etc.

[0103] S32. Construct a crop growth model based on the collected data, and determine the water requirement of crops in the arid area within a preset time through the constructed crop growth model.

[0104] Among them, a crop growth model is constructed based on the collected data, and the water requirement of crops in arid areas within a preset time is determined by the constructed crop growth model, including:

[0105] S321. Construct photosynthesis simulation model based on crop variety data and meteorological data in arid areas;

[0106] The photosynthesis simulation model simulates photosynthesis based on the crop variety data in the orchard and the meteorological data of the area where the orchard is located, such as light intensity, temperature, carbon dioxide concentration and other factors. For example, in a day with sufficient light, the photosynthesis rate is calculated based on the absorption and conversion capacity of the photosynthetic pigments of the leaves and the effect of temperature on the activity of photosynthetic enzymes. The expression is as follows:

[0107]

[0108] Where: P n is the net photosynthetic rate, P max is the maximum photosynthetic rate, I is the light intensity, K is the light intensity corresponding to the light saturation point, T is the actual temperature, T min 、T max and T opt They are the minimum temperature, maximum temperature and optimum temperature for photosynthesis respectively.

[0109] S322. Constructing a respiration simulation model based on crop variety data and meteorological data in arid areas;

[0110] The respiration simulation model simulates the respiration of crops based on the variety data of crops in the orchard and the meteorological data of the area where the orchard is located, taking into account the maintenance respiration and growth respiration of crops. Maintenance respiration is the basic respiration to maintain the life activities of crop cells, and growth respiration is the respiration used for the growth of new tissues. The expression is as follows:

[0111]

[0112] Where: R is the respiratory rate, R m is the basal respiration rate at the reference temperature, T ref is the reference temperature.

[0113] S323. Construct a transpiration simulation model based on crop variety data, soil texture data and meteorological data in arid areas;

[0114] The transpiration simulation model simulates the transpiration of water demand based on the crop variety data in the orchard, the soil texture data of the orchard, and the meteorological data of the area where the orchard is located. The model calculates the transpiration rate through the stomatal conductance of the leaves, which is affected by multiple factors such as light, temperature, humidity, and carbon dioxide concentration. The expression is as follows:

[0115]

[0116] Where: g s is the stomatal conductance, g s0 is the initial stomatal conductance, m is the sensitivity coefficient of stomatal to changes in photosynthetic rate and intercellular carbon dioxide concentration, A is the photosynthetic rate, C s is the intercellular carbon dioxide concentration.

[0117] Then the crop transpiration water volume is calculated according to the transpiration rate, that is, the transpiration rate multiplied by the leaf area index and the time step can get the transpiration water volume per unit land area. The calculation formula is as follows:

[0118] ET=E×LAI×Δt

[0119] Where: ET is the transpiration water per unit land area, E is the transpiration rate, LAI is the leaf area index, and Δt is the time step.

[0120] S324. Determine a crop growth model according to the constructed photosynthesis simulation model, respiration simulation model and transpiration simulation model, and determine the water requirements of crops in arid areas at different growth stages within a preset time based on the crop growth model.

[0121] The growth status of crops is updated according to the distribution of photosynthetic products produced by photosynthesis to various organs, such as leaves, stems, and fruits. For example, part of the photosynthetic products are used for leaf growth, increasing the leaf area index; part is used for fruit expansion, increasing the volume and weight of the fruit. At the same time, the root system will also grow and develop according to the water and nutrient conditions in the soil and the growth needs of the aboveground parts. The model will update parameters such as the distribution and absorption capacity of the root system.

[0122] Finally, the model outputs the water requirements of orchard crops at different growth stages, such as germination, flowering, fruit expansion, and maturity, every day or at each time step. In addition to water requirements, other related crop growth indicators are also output, such as leaf area index change curve, aboveground biomass accumulation, root growth dynamics, etc., to assist in the comprehensive assessment of the growth status of orchard crops and their response to the environment.

[0123] The above-mentioned embodiment can fully simulate the physiological processes of orchard crops at different growth stages, including photosynthesis, respiration and transpiration, accurately calculate the changes in the water demand of crops, and provide an accurate data basis for subsequent intelligent irrigation.

[0124] Embodiment 4

[0125] See also Figure 4 In the present invention, the remote control terminal is used to analyze the analysis results transmitted by the cloud platform, determine the irrigation strategy corresponding to the drought area, and control the irrigation device to perform intelligent irrigation through the determined irrigation strategy, including the following steps:

[0126] S41. The remote control terminal determines the precipitation within a preset time according to the weather environment information collected in the orchard area;

[0127] Weather and environmental information in the orchard area can be obtained through multiple channels. First, based on observations at meteorological stations, rain gauges are used to measure and obtain data after calibration and quality control. Second, with the help of satellite remote sensing, based on the principles of microwave and infrared sensors, precipitation is estimated through data processing and algorithm optimization combined with ground calibration. Third, numerical weather forecast models are used to solve equations based on atmospheric physics principles, and precipitation is predicted after verification and uncertainty assessment. The specific data collection method is not limited here.

[0128] S42. Calculate the target water requirement of crops in the arid area within the preset time based on the determined precipitation and the water requirement of crops in the arid area within the preset time;

[0129] 1. Calculate and determine the difference between precipitation and the water requirement of crops in arid areas within a preset time;

[0130] When the preset time here is the daily scale time, the difference between the crop water demand and precipitation is calculated on a one-day time scale. If the precipitation is greater than the water demand, theoretically no irrigation is needed on that day, and drainage measures may even need to be considered to avoid soil waterlogging. If the precipitation is less than the water demand, the difference is the amount of water that needs to be supplemented by the irrigation device on that day. When the preset time here is the weekly scale and growth stage, the difference between the cumulative water demand and precipitation is calculated on this scale. Because even if the precipitation is greater than the water demand on a certain day, there may still be a general lack of water within a week. At the same time, the water demand of crops at different growth stages varies greatly, and the water demand of crops usually increases significantly during the fruit expansion period. Combined with the water demand characteristics of the crop growth stage, analyze the water balance during this period to determine the overall irrigation demand.

[0131] 2. When the calculated difference meets the conditions, the target amount of water to be supplemented by the irrigation device is determined based on the water holding capacity of the soil in the arid area and the irrigation efficiency of the irrigation device.

[0132] When the calculated difference satisfies the condition that the precipitation is less than the water requirement, the target amount of water to be supplemented by the irrigation device needs to be determined based on the water holding capacity of the soil in the arid area and the irrigation efficiency of the irrigation device. Different types of soil have different water holding capacities. The field water holding capacity and wilting coefficient of the soil need to be determined according to the texture type of the orchard soil. When the soil moisture content is lower than the wilting coefficient, the crop will be subjected to drought stress; when the soil moisture content reaches the field water holding capacity, the pore water in the soil reaches a saturated state. Calculate the difference between the current soil moisture content and the field water holding capacity, and combine it with the difference between the water requirement and precipitation analyzed above to more accurately determine the amount of water that needs to be input by the irrigation device. The efficiency of the irrigation device itself will also affect the actual amount of water that needs to be input. According to the type of irrigation device and its previous operation, estimate its irrigation efficiency. If the irrigation efficiency is 80%, then when calculating the amount of water that needs to be input, this factor should be considered and the amount of water should be appropriately increased to ensure that the crop can get enough water.

[0133] S43. constructing an objective function with the shortest irrigation time and the minimum irrigation amount as the target parameters;

[0134]

[0135] Where: n is the irrigation unit, T is the total irrigation time of the irrigation device, t i is the irrigation time of the i-th irrigation unit, Q is the total irrigation amount of the irrigation device, qi is the irrigation amount of the ith irrigation unit, Z is the total objective function value, w1 and w2 are the weight coefficients of T and Q respectively, and w1+w2=1.

[0136] Minimizing the total irrigation time means shortening the overall irrigation time by rationally arranging the working hours of each irrigation unit; the total irrigation amount is minimized by optimizing the selection and controlling the irrigation amount of each irrigation unit, thereby achieving the purpose of saving water.

[0137] S44. Determine constraints, which include constraints on target water demand for crops in arid areas, constraints on irrigation unit performance, constraints on the number of irrigation units working simultaneously, constraints on water source supply capacity, and constraints on soil water holding capacity;

[0138] Meeting target water demand constraints for crops in arid areas:

[0139]

[0140] Where: Q need The amount of water required by crops in arid areas within a preset time.

[0141] This is the basic condition to ensure the normal growth of crops. Any feasible irrigation plan must ensure that the irrigation volume is sufficient to meet the water requirements for crop growth.

[0142] Irrigation unit performance constraints:

[0143] q i,min ≤q i ≤q i,max

[0144] t i，mmin ≤t i ≤t i，max

[0145] Where: q i,mmin and q i,mmax are the minimum and maximum allowable irrigation amounts for the ith irrigation unit, t i，mmin and t i，max are the shortest and longest irrigation time of the i-th irrigation unit, respectively.

[0146] The above limitations are due to the equipment characteristics of the irrigation unit. For example, the flow range of the sprinkler head determines the upper and lower limits of its irrigation volume, while factors such as the operating stability of the equipment limit its maximum and minimum continuous working time.

[0147] Constraints on the number of irrigation units working simultaneously:

[0148]

[0149] Where: m is the maximum number of irrigation units allowed to work simultaneously, xi is a 0-1 variable, x i =1 means the i-th irrigation unit is working, x i =0 means not working.

[0150] This constraint ensures that the number of irrigation units that are turned on at the same time does not exceed the maximum value that the system can withstand.

[0151] Constraints on water supply capacity:

[0152]

[0153] Where: Q source The maximum water supply per hour of the water source.

[0154] This constraint ensures that the water source can continuously and stably supply water to the irrigation system during the irrigation process, and there will be no water shortage.

[0155] Soil water holding capacity constraints:

[0156]

[0157] Where: Q initial is the initial soil moisture content, Q soil The maximum water holding capacity of the soil.

[0158] The water holding capacity of soil has an upper limit. Excessive irrigation can lead to adverse consequences such as waterlogging in the soil. This constraint can be used to avoid the problem of soil quality deterioration and water resource waste caused by improper irrigation.

[0159] S45. Determine the irrigation strategy corresponding to each drought area based on the objective function and constraint conditions.

[0160] According to the optimization problem constructed based on the above objective function and constraints, a suitable optimization algorithm is used here to solve it. There is no limitation on the linear or nonlinear complexity of the model, so there is no limitation on the optimization algorithm here. After solving it through the selected optimization algorithm, the result obtained will clarify the irrigation time t of each irrigation unit. i and irrigation volume q i The value of x i The value of determines which irrigation units participate in the work, which constitutes the optimal irrigation plan. i =1, the number of irrigation units corresponding to the optimal solution can be obtained.

[0161] Finally, after the remote control end determines the irrigation plan, the remote control end generates an irrigation instruction according to the determined irrigation plan, and instructs the irrigation unit number, irrigation start time, irrigation duration, irrigation amount, irrigation frequency and other information. Through the established communication connection, the remote control end sends the irrigation instruction to the irrigation device. After the irrigation device receives the irrigation instruction, the internal control unit parses the instruction and converts the instruction into an operation signal of the irrigation device according to the parameters in the instruction. During the process of the irrigation device executing the irrigation instruction, the sensor on the irrigation device will monitor the operating status of the irrigation device in real time. If the remote control end finds that the actual irrigation situation does not match the predetermined plan, such as abnormal flow, irrigation time deviation, etc., it will make adjustments based on the feedback data. During the entire irrigation process, the remote control end will record the relevant irrigation data, including the time, location, irrigation unit, irrigation amount, weather conditions and other information of each irrigation. By analyzing the irrigation record data, the effect of the irrigation plan is evaluated to optimize the irrigation strategy.

[0162] In the above embodiment, the irrigation device is remotely controlled by the remote control terminal. The orchard manager only needs to remotely control it on the mobile phone or computer terminal, and does not need to go to the site to manually operate the irrigation equipment. The automated intelligent irrigation system reduces the dependence on manual irrigation and reduces labor costs. The optimal irrigation plan determined by the remote control terminal can not only consider the irrigation amount, but also combine factors such as irrigation time and the operating efficiency of the irrigation device to reasonably arrange the working time and working mode of the irrigation unit, thereby effectively reducing water resource waste and energy consumption.

[0163] It is understandable that those skilled in the art can, under the guidance of the above embodiments, combine various implementation methods in the above embodiments to obtain technical solutions of multiple implementation methods.

[0164] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent irrigation system for orchards in arid areas, characterized in that: include: An orchard image data collection terminal, a crop growth data collection terminal, a cloud platform, a remote control terminal and an irrigation device, wherein the cloud platform is respectively connected to the orchard image data collection terminal, the crop growth data collection terminal and the remote control terminal, and the remote control terminal is connected to the irrigation device; The orchard image data acquisition terminal uses a drone to take real-time image information of crops in the orchard, and uploads the captured images to the cloud platform; The crop growth data collection terminal collects multi-dimensional growth data of crops through sensors, and uploads the collected data to the cloud platform; The cloud platform includes a drought area monitoring module and a crop water requirement prediction module. The drought area monitoring module is used to perform image recognition on image information of crops in an orchard to determine potential drought areas in the orchard; the crop water requirement prediction module is used to analyze and predict the growth data of crops in drought areas to obtain the water requirement of crops in drought areas within a preset time; the cloud platform transmits the analysis results generated by the drought area monitoring module and the crop water requirement prediction module to the remote control terminal; The remote control terminal is used to analyze the analysis results transmitted by the cloud platform, determine the irrigation strategy corresponding to the drought area, and control the irrigation device to perform intelligent irrigation through the determined irrigation strategy.

2. The intelligent irrigation system for orchards in arid areas according to claim 1, characterized in that: The working method of the arid area monitoring module includes: Grid division is performed based on the received image information, and data corresponding to each divided grid is preprocessed; The preprocessed data is input into a preset orchard convolutional neural network, and a probability vector of each grid is output, wherein the probability vector includes drought elements and non-drought elements; The distribution of all potential drought areas in the orchard is determined according to the probability vector.

3. The intelligent irrigation system for orchards in arid areas according to claim 2, characterized in that: The training process of the preset orchard convolutional neural network includes the following: Collect historical orchard images taken by drones, which include information on the drought and non-drought states of orchard conditions in different seasons, at different times and under different weather conditions; The collected historical orchard images are preprocessed and the preprocessed data is input into the constructed model, which includes the first convolutional layer, the second convolutional layer, the pooling layer, the first fully connected layer and the second fully connected layer; The binary cross entropy loss function and optimizer are used for multiple iterative training. When the preset number of training rounds is reached, the model training is completed and the output is a probability vector containing arid areas and non-drought areas.

4. The intelligent irrigation system for orchards in arid areas according to claim 1, characterized in that: The working method of the crop water requirement prediction module includes: Collect data on crop varieties, soil texture and meteorological data in arid areas; A crop growth model is constructed based on the collected data, and the water requirement of crops in arid areas within a preset time is determined through the constructed crop growth model.

5. The intelligent irrigation system for orchards in arid areas according to claim 4, characterized in that: The method of constructing a crop growth model based on the collected data and determining the water requirement of crops in arid areas within a preset time by using the constructed crop growth model includes: Construct photosynthesis simulation model based on crop variety data and meteorological data in arid areas; Construct a respiration simulation model based on crop variety data and meteorological data in arid areas; Construct a transpiration simulation model based on crop variety data, soil texture data and meteorological data in arid areas; A crop growth model is determined according to the constructed photosynthesis simulation model, respiration simulation model and transpiration simulation model, and the water requirement of crops in arid areas at different growth stages within a preset time is determined based on the crop growth model.

6. The intelligent irrigation system for orchards in arid areas according to claim 5, characterized in that: The expressions of the photosynthesis simulation model, the respiration simulation model and the transpiration simulation model include: Photosynthesis simulation model: Where: P n is the net photosynthetic rate, P max is the maximum photosynthetic rate, I is the light intensity, K is the light intensity corresponding to the light saturation point, T is the actual temperature, T min , T max and T opt They are the minimum temperature, maximum temperature and optimum temperature for photosynthesis respectively; Respiration simulation model: Where: R is the respiratory rate, R m is the basal respiration rate at the reference temperature, T ref is the reference temperature; Transpiration simulation model: Where: g s is the stomatal conductance, g s0 is the initial stomatal conductance, m is the sensitivity coefficient of stomatal to changes in photosynthetic rate and intercellular carbon dioxide concentration, A is the photosynthetic rate, C s is the intercellular carbon dioxide concentration.

7. The intelligent irrigation system for orchards in arid areas according to claim 1, characterized in that: The remote control terminal is used to analyze the analysis results transmitted by the cloud platform, determine the irrigation strategy corresponding to the drought area, and control the irrigation device to perform intelligent irrigation according to the determined irrigation strategy, including: The remote control terminal determines the precipitation within a preset time according to the collected weather environment information of the area where the orchard is located; Calculate the target water requirement of the crops in the arid area within a preset time based on the determined precipitation and the water requirement of the crops in the arid area within a preset time; The objective function is constructed with the shortest irrigation time and the minimum irrigation amount as the target parameters; Determining constraints, wherein the constraints include a target water demand constraint for crops in arid areas, an irrigation unit performance constraint, a simultaneous operation quantity constraint for irrigation units, a water source water supply capacity constraint, and a soil water holding capacity constraint; The irrigation strategy corresponding to each arid area is determined according to the objective function and the constraint conditions.

8. The intelligent irrigation system for orchards in arid areas according to claim 7, characterized in that: The method of calculating the target water requirement of the crops in the arid area within the preset time based on the determined precipitation and the water requirement of the crops in the arid area within the preset time includes: Calculate and determine the difference between precipitation and the water requirement of crops in arid areas within a preset time; When the calculated difference meets the conditions, the target amount of water to be supplemented by the irrigation device is determined based on the water holding capacity of the soil in the arid area and the irrigation efficiency of the irrigation device.

9. The intelligent irrigation system for orchards in arid areas according to claim 7, characterized in that: The objective function is constructed by taking the shortest irrigation time and the minimum irrigation amount as the objective parameters, including: Where: n is the irrigation unit, T is the total irrigation time of the irrigation device, t i is the irrigation time of the i-th irrigation unit, Q is the total irrigation amount of the irrigation device, q i is the irrigation amount of the ith irrigation unit, Z is the total objective function value, w1 and w2 are the weight coefficients of T and Q respectively, and w1+w2=1.

10. The intelligent irrigation system for orchards in arid areas according to claim 7, characterized in that: The determined constraints include the target water demand constraint of crops in arid areas, the irrigation unit performance constraint, the number of irrigation units working simultaneously constraint, the water source water supply capacity constraint and the soil water holding capacity constraint, including: Meeting target water demand constraints for crops in arid areas: Where: Q need The amount of water required for crops in dry areas within a preset time; Irrigation unit performance constraints: q i,min ≤q i ≤q i,max t i,min ≤t i ≤t i,max Where: q i,min and q i,max are the minimum and maximum allowable irrigation amounts for the ith irrigation unit, t i,min and t i,max are the shortest and longest irrigation time of the ith irrigation unit, respectively; Constraints on the number of irrigation units working simultaneously: Where: m is the maximum number of irrigation units allowed to work simultaneously, x i is a 0-1 variable, x i =1 means the i-th irrigation unit is working, x i =0 means not working; Constraints on water supply capacity: Where: Q source The maximum water supply per hour for the water source; Soil water holding capacity constraints: Where: Q inital is the initial soil moisture content, Q soil The maximum water holding capacity of the soil.

Images