Intelligent control system and method for fruit tree irrigation

Abstract

The invention discloses an intelligent control system and method for fruit tree irrigation, and the system comprises a soil humidity detection module, a fruit tree information database module, an irrigation decision module and an irrigation execution module, or comprises an image collection module, an environment monitoring module, a fruit tree growth analysis module, an irrigation decision module and an irrigation execution module. According to the method, a more scientific and effective auxiliary means is provided for fruit tree pruning, and the yield and quality of fruit tree cultivation can be improved.

Description

Technical Field

[0001] The invention relates to the technical field of fruit tree cultivation, and in particular to an intelligent control system and method for fruit tree irrigation. Background Art

[0002] Reasonable irrigation is one of the key factors for fruit tree growth and high yield. However, traditional fruit tree irrigation methods have many problems.

[0003] On the one hand, irrigation decisions often lack scientificity and precision. Fruit farmers usually irrigate based on experience and fixed irrigation schedules, without fully considering the differences in various factors such as fruit tree varieties, tree ages, soil types, soil moisture, climatic conditions, and fruit tree growth stages. This may lead to insufficient or excessive irrigation, affecting the growth of fruit trees and fruit quality, and may even cause problems such as water waste and soil erosion.

[0004] On the other hand, the control of the irrigation process is not precise enough. Traditional irrigation methods make it difficult to achieve precise control of irrigation volume, irrigation time and irrigation uniformity, and cannot meet the refined water needs of fruit trees.

[0005] With the development of modern information technology and agricultural technology, new ideas and methods have been provided for the intelligent control of fruit tree irrigation. Summary of the invention

[0006] The present invention provides an intelligent control system and method for fruit tree irrigation based on a soil moisture model and fruit tree water demand analysis, the system comprising:

[0007] Soil moisture detection module: Through soil moisture sensors and data analysis technology, the moisture status of the orchard soil is monitored in real time, including moisture data of soil layers at different depths.

[0008] Fruit tree information database module: stores biological characteristics of different fruit tree varieties, such as growth cycle, water demand characteristics at different growth stages, root distribution characteristics, etc., as well as basic information of each fruit tree in the orchard, such as tree age, variety, planting location, etc.

[0009] Irrigation decision module: Combine soil moisture detection data with data in the fruit tree information database, use the established soil moisture model and fruit tree water requirement analysis model to develop a personalized irrigation plan for each fruit tree, including irrigation amount, irrigation time and irrigation location.

[0010] Irrigation execution module: According to the irrigation plan formulated by the irrigation decision module, the irrigation equipment is precisely controlled to perform irrigation operations to ensure the accuracy and uniformity of irrigation.

[0011] The irrigation control method is:

[0012] 1. Soil moisture detection: The soil moisture detection module can be used to monitor the moisture status of the orchard soil in real time.

[0013] 2. Fruit tree information acquisition: Obtain the basic information and variety biological characteristics information of each fruit tree in the orchard from the fruit tree information database module.

[0014] 3. Irrigation plan formulation: Combine soil moisture detection data and fruit tree information, and use the irrigation decision module to formulate personalized irrigation plans for each fruit tree.

[0015] 4. Irrigation execution: According to the irrigation plan, the irrigation execution module controls the irrigation equipment to perform irrigation operations.

[0016] On the other hand, the present invention provides another fruit tree irrigation intelligent control system and method based on image recognition and environmental factor integration, the system comprising:

[0017] Image acquisition module: Install high-definition cameras or drone-mounted cameras in the orchard to regularly collect image information of fruit trees, including the crown morphology, leaf color and size, fruit development, etc.

[0018] Environmental monitoring module: Set up environmental sensors such as temperature, humidity, light intensity, wind direction and speed to monitor the environmental conditions of the orchard in real time.

[0019] Fruit tree growth analysis module: Use machine learning algorithms to analyze collected images and environmental data, identify the growth stage and growth status of fruit trees, and determine the water requirements of fruit trees.

[0020] Irrigation decision module: According to the analysis results of the fruit tree growth analysis module, combined with the characteristics of fruit tree varieties and environmental conditions, formulate irrigation plans, including irrigation amount, irrigation time and irrigation location.

[0021] Irrigation execution module: controls irrigation equipment to perform irrigation operations according to the irrigation plan formulated by the irrigation decision module.

[0022] The irrigation control method is:

[0023] 1. Image acquisition and environmental monitoring: Obtain fruit tree images and environmental data through the image acquisition module and environmental monitoring module.

[0024] 2. Fruit tree growth analysis: Input the collected images and environmental data into the fruit tree growth analysis module, and use the trained machine learning algorithm to analyze the growth stage and growth status of the fruit trees to determine the water requirements of the fruit trees.

[0025] 3. Irrigation plan formulation: Based on the results of fruit tree growth analysis, combined with the characteristics of fruit tree varieties and environmental conditions, the irrigation decision module formulates an irrigation plan.

[0026] 4. Irrigation execution: According to the irrigation plan, the irrigation execution module controls the irrigation equipment to perform irrigation operations.

[0027] Beneficial effects:

[0028] It can accurately analyze soil moisture conditions and water requirements of fruit trees, providing a scientific basis for formulating precise irrigation plans; it takes into account individual differences in fruit trees and soil moisture differences, achieves precise irrigation, and improves irrigation effects and water resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 , schematic diagram of the process flow of the invention. DETAILED DESCRIPTION

[0030] Embodiment 1:

[0031] 1. Implementation of soil moisture detection module

[0032] Reasonably arrange soil moisture sensors in the orchard and install them at a certain depth and interval to ensure that the soil moisture conditions can be fully and accurately detected. Regularly calibrate and maintain the sensors, and use data analysis technology to process the data collected by the sensors to obtain more accurate soil moisture data.

[0033] 2. Implementation of fruit tree information database module

[0034] Collect information on the biological characteristics of different fruit tree varieties and basic information about each fruit tree in the orchard, enter this information into the database, and update it regularly to ensure the accuracy and completeness of the data.

[0035] 3. Implementation of irrigation decision module

[0036] ·Write an irrigation decision-making program based on the soil moisture detection data and the data in the fruit tree information database, and use the established soil moisture model and fruit tree water requirement analysis model to calculate the irrigation plan for each fruit tree.

[0037] 4. Implementation of irrigation execution module

[0038] ·Control the irrigation equipment to perform irrigation operations according to the irrigation plan. The irrigation equipment should have

[0039] The function of accurately controlling the irrigation amount and irrigation location, such as using drip irrigation system, micro sprinkler irrigation system, etc.

[0041] Innovation 1: Construction of soil moisture model and fruit tree water demand analysis model

[0042] Mathematical model algorithm:

[0043] Soil moisture model: Assume that soil moisture index is H = (h1, h2, ..., h n ), where h i Represents the i-th humidity index (such as surface soil humidity is h1, middle soil humidity is h2, etc.). A soil humidity dynamic change model is established, taking into account the influence of factors such as soil humidity change law, precipitation input, evaporation and transpiration loss, and fruit tree absorption on humidity. For example, for surface soil humidity, its dynamic change model can be expressed as: Among them, p1 is the precipitation input, e1 is the evapotranspiration, and u1 is the absorption of fruit trees.

[0044] Fruit tree water demand analysis model: According to the different water demand of fruit trees at different growth stages (such as budding period, flowering period, fruit expansion period, maturity period, etc.), a fruit tree water demand analysis model is established. Assume that the water demand of fruit trees at a certain growth stage t is d tj Through the analysis of a large amount of experimental data and the growth law of fruit trees, we determined d tj The relationship between the fruit tree variety, tree age, soil moisture and other factors. For example, the water demand of a certain variety of fruit trees during the fruit expansion period is d t1 There may be the following relationship between tree age a and soil moisture h1: t1 =β0+β1a+β2h1, where β0, β1, β2 are coefficients determined by experiments.

[0045] Advantages: It can accurately analyze soil moisture conditions and water requirements of fruit trees, providing a basis for formulating precise

[0046] Provide scientific basis for standard irrigation scheme.

[0047] Innovation 2: Development and implementation of personalized irrigation plans

[0048] Irrigation plan formulation

[0049] · According to the calculation results of soil moisture model and fruit tree water demand analysis model, a personalized irrigation plan is formulated for each fruit tree.

[0050] If the soil moisture is low and the water demand is high during the fruit expansion period, the irrigation decision module will formulate plans to increase the irrigation amount, irrigate specific areas around the roots of the fruit trees, and irrigate at appropriate times (such as morning or evening).

[0051] Advantages: It takes into account the individual differences of fruit trees and the differences in soil moisture, realizes precise irrigation, and improves irrigation effect and water resource utilization.

[0052] Embodiment 2:

[0053] 1. Image acquisition module implementation

[0054] Reasonably arrange high-definition cameras in the orchard or arrange drones for regular photography to ensure that the orchard is fully covered and various image information of the fruit trees is collected. The camera or camera should have sufficient resolution and clarity to collect high-quality images.

[0055] 2. Implementation of environmental monitoring module

[0056] Evenly distribute temperature and repeatability in the orchard, which can be optimized as follows: evenly distribute environmental sensors such as temperature, humidity, light intensity, wind direction and speed in the orchard to ensure the accuracy and reliability of data. Regularly calibrate and maintain the sensors.

[0057] 3. Implementation of fruit tree growth analysis module

[0058] Collect a large number of annotated fruit tree images and environmental data to build a training dataset. Use a deep learning framework (such as TensorFlow or PyTorch) to train the convolutional neural network model, and set appropriate training parameters such as learning rate and batch size until the model achieves satisfactory accuracy and recall.

[0059] 4. Implementation of irrigation decision module

[0060] ·Write irrigation decision-making procedures and formulate irrigation plans based on the results of fruit tree growth analysis, combined with fruit tree variety characteristics and environmental conditions.

[0061] 5. Implementation of irrigation execution module

[0062] ·Control the irrigation equipment to carry out irrigation operations according to the irrigation plan. The irrigation equipment should have

[0063] The function of accurately controlling the irrigation amount and irrigation location, such as using drip irrigation system, micro sprinkler irrigation system, etc.

[0065] Innovation 1: Image recognition and machine learning algorithm application

[0066] Mathematical model algorithm

[0067] Construct a fruit tree growth image recognition model. Use the convolutional neural network (CNN) algorithm to train a large number of labeled fruit tree images.

[0068] Suppose the input image is I. After operations such as convolutional layer and pooling layer, the output is the predicted value of the growth stage S and growth status G (such as health, water shortage, nutrient deficiency, etc.) of the fruit tree.

[0069] By continuously adjusting the network parameters, the accuracy and recall of the model are optimized. The cross entropy loss function is used. where y i is the true label, To predict the value, the network weights are adjusted through the back-propagation algorithm to minimize the loss function.

[0070] Advantages: It can accurately identify the growth stage and growth status of fruit trees, and provide guidance for formulating irrigation plans.

[0071] The plan provides an accurate basis.

[0072] Innovation point 2: Comprehensive consideration of environmental factors and fruit tree growth

[0073] Data analysis methods

[0074] Collect a large amount of historical fruit tree growth data and corresponding environmental data, and establish an environment-fruit tree growth correlation database.

[0075] The relationship between environmental factors and fruit tree growth was analyzed using multiple linear regression analysis. Let the fruit tree growth index be Y, and the environmental factors be X1, X2, …, X n (such as temperature is X1, humidity is X2, etc.), establish a regression model

[0076] Y=β0+β1X1+β2X2+…+β n X n , the regression coefficients β0, β1, …, β are estimated by the least squares method n .

[0077] Advantages: Considering the impact of environmental factors on the growth of fruit trees, it is possible to more accurately determine the water requirements of fruit trees and develop more reasonable irrigation plans.

[0078] Example 3

[0079] Comparative synergistic test

[0080] (I) Experimental setup

[0081] 1. Experimental area selection: Four planting areas with different environmental conditions were selected as experimental sites, marked as Area A, Area B, Area C and Area D. Area A has high temperature and high humidity climate conditions, high soil fertility and moderate pH; Area B has a cooler and drier climate, moderate soil fertility and is slightly acidic; Area C is in a cold and humid environment, with low soil fertility and is slightly alkaline; Area D has special microclimate and soil conditions, such as large local temperature fluctuations and specific mineral components in the soil.

[0082] 2. Variety selection: 12 common fruit tree varieties were selected as test objects, namely apple, pear, peach, apricot, plum, cherry, grape, citrus, banana, mango, longan and lychee.

[0083] 3. Experimental grouping: In each experimental area, 12 fruit tree varieties were divided into three groups, and irrigation control was performed using the two methods of the present invention (based on soil moisture model and fruit tree water demand analysis, based on image recognition and environmental factor integration), and the number of varieties controlled by irrigation in each group was 4. Another group was set as a control group, using the traditional irrigation method, and the number of varieties controlled by irrigation was also 4 from the 12 varieties.

[0084] 2. Evaluation indicators

[0085] 1. Irrigation effect evaluation: The irrigation effect is evaluated by measuring indicators such as the uniformity of soil moisture, the moisture content of fruit tree leaves, and the size and quality of the fruit, and a comprehensive score is given, with a full score of 100.

[0086] 2. Water resource utilization rate: Calculate the proportion of irrigation water absorbed and utilized by fruit trees by measuring the residual moisture content in the soil and the moisture content absorbed by the fruit trees.

[0087] 3. Fruit tree yield: The average yield of each variety in each test area is calculated in kilograms. The planting cycle is the time from planting to fruit maturity.

[0088] (III) Test results

[0089] The test results are shown in the following table:

[0090] Region A test results

[0091]

[0092] Region B test results

[0093]

[0094] Region C test results

[0095]

[0096]

[0097] Region D test results

[0098]

[0099]

[0100] in conclusion:

[0101] 1. In terms of irrigation effect evaluation, the method based on soil moisture model and fruit tree water demand analysis and the method based on image recognition and environmental factor integration were significantly higher than the traditional method in the tests in four different environmental regions. This shows that the two methods of the present invention can more effectively improve the irrigation effect, make the soil moisture more uniform, the moisture content of fruit tree leaves more appropriate, and the fruit quality better.

[0102] 2. In terms of water resource utilization, the water resource utilization of the two inventive methods is significantly higher than that of the traditional method, which shows that the method of the present invention can more effectively utilize water resources and reduce water resource waste.

[0103] 3. In terms of fruit tree yield, the average yield of the varieties irrigated and controlled by the two inventive methods is significantly higher than that of the varieties irrigated and controlled by the traditional method, which shows that the method of the present invention has a positive impact on the yield of fruit trees.

[0104] 4. Overall, the test results of the method based on image recognition and environmental factors are slightly better than those based on soil moisture model and fruit tree water demand analysis in most areas, but both methods are significantly better than the traditional method. The two methods of the present invention provide more scientific and effective control means for fruit tree irrigation, which are helpful to improve the yield and quality of fruit tree cultivation and improve the utilization rate of water resources.

Claims

1. An intelligent control system for fruit tree irrigation, characterized in that: The system includes a soil moisture detection module, a fruit tree information database module, an irrigation decision module and an irrigation execution module, or includes an image acquisition module, an environment monitoring module, a fruit tree growth analysis module, an irrigation decision module and an irrigation execution module.

2. The fruit tree irrigation intelligent control system according to claim 1, characterized in that: The soil moisture detection module detects soil moisture conditions, the fruit tree information database module stores fruit tree information, and the irrigation decision module formulates irrigation plans.

3. The fruit tree irrigation intelligent control method according to claim 1, characterized in that Including the construction of soil moisture model and fruit tree water demand analysis model, and the construction of fruit tree growth image recognition model. The convolutional neural network (CNN) algorithm is used to train a large number of labeled fruit tree images. Let the input image be I. After operations such as convolution layer and pooling layer, the output is the predicted value of fruit tree growth stage S and growth state G (such as health, water shortage, nutrient deficiency, etc.). By continuously adjusting the network parameters, the accuracy and recall of the model are optimized. The cross entropy loss function is used. where y i is the true label, To predict the value, the network weights are adjusted through the back-propagation algorithm to minimize the loss function.

4. The fruit tree irrigation intelligent control system according to claim 1 is characterized in that The image acquisition module collects data, the environment monitoring module, the fruit tree growth analysis module, the irrigation decision module formulates irrigation plans, and the irrigation execution module.

5. The fruit tree irrigation intelligent control method according to claim 1, characterized in that It includes the application of image recognition and machine learning algorithms, as well as comprehensive consideration of environmental factors and fruit tree growth.

6. The fruit tree irrigation intelligent control method according to claim 1, characterized in that The specific steps are soil moisture detection, fruit tree information acquisition, irrigation plan formulation, and irrigation implementation. Soil moisture model: Assume that the soil moisture index is H = (h1, h2, ..., h n ), where hi represents the i-th humidity index (such as surface soil humidity is h1, middle soil humidity is h2, etc.). A soil humidity dynamic change model is established, taking into account the influence of factors such as soil humidity change law, precipitation input, evaporation and transpiration loss, and fruit tree absorption on humidity. For example, for surface soil humidity, its dynamic change model can be expressed as: Among them, p1 is the precipitation input, e1 is the evapotranspiration, and u1 is the absorption of fruit trees. Fruit tree water demand analysis model: According to the different water demand of fruit trees at different growth stages (such as budding period, flowering period, fruit expansion period, maturity period, etc.), a fruit tree water demand analysis model is established. Assume that the water demand of fruit trees at a certain growth stage t is d tj Through the analysis of a large amount of experimental data and the growth law of fruit trees, we determined d tj The relationship between the fruit tree variety, tree age, soil moisture and other factors. For example, the water demand of a certain variety of fruit trees during the fruit expansion period is d t1 There may be the following relationship with tree age a and soil moisture h1: d t1 =β0+β1a+β2h1, where β0, β1, β2 are coefficients determined by experiments.

Images