Wine grape water and fertilizer control method, device and system and storage medium
Through the deep neural network model combining the generative adversarial network and the selective nuclear convolutional neural network, precise water fertilizer control of wine grapes is achieved, and the problems of waste of water resources and low fertilizer utilization in traditional methods are solved, and automated fertilization and high-efficiency water are realized. Resource management.
Patent Information
- Application Number
- CN202411078006.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-07
- Publication Date
- 2024-12-06
AI Technical Summary
The traditional water and fertilizer control method of wine grapes depends on feelings and experience, and cannot accurately grasp the amount and time of fertilizer application, resulting in waste of water resources and low fertilizer utilization, and consume a lot of manpower.
The deep neural network model is adopted, combining the generative adversarial network and the selective nuclear convolutional neural network, training is carried out based on historical soil and meteorological data, data is collected in real time for fertilization and irrigation control, and the fertilization and irrigation scheme is optimized using time iteration algorithms.
It realizes automated fertilization, saves labor costs, improves fertilizer utilization, and reduces water resource waste.
Smart Images
Figure CN119090665A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of water and fertilizer management of wine grapes, and in particular relates to a water and fertilizer control method and device, a system, and a storage medium for wine grapes. Background Art
[0002] Wine grapes refer to grape varieties whose main production purpose is to make wine. They can be roughly divided into red varieties, white varieties, etc. There are many varieties of wine grapes, which are very different from the grapes we usually eat. Wine grapes have thick skins, less flesh, more juice, and small grains, which are basically not suitable for consumption. The fruits of wine grapes are smaller in size than fresh-table grapes, and the skin is also thicker. The main wine grape producing areas in my country are distributed in North China, Bohai Bay, Xinjiang, Gansu, Ningxia and other places. In the traditional water and fertilizer control of wine grapes, watering and fertilizing are mostly based on feelings or experience. The appropriate amount and time of fertilization cannot be mastered, which cannot meet the needs of crops. In addition, manual fertilization consumes manpower, has low water utilization rate, and wastes water resources. Summary of the invention
[0003] The technical problem to be solved by the present invention is to provide a method and device, a system and a storage medium for controlling water and fertilizer of wine grapes.
[0004] To achieve the above object, the present invention adopts the following technical solution:
[0005] A method for controlling water and fertilizer of wine grapes, comprising:
[0006] Step S1, obtaining historical soil and meteorological data of the grape field;
[0007] Step S2: training a deep neural network model based on historical soil and meteorological data;
[0008] Step S3: input the soil and meteorological data of the grape field collected in real time into the trained deep neural network model to achieve fertilization and irrigation of the grape field.
[0009] Preferably, the deep neural network model is a multidimensional parameter prediction model based on a mixture of a generative adversarial network DD-GAN and a selective kernel convolutional neural network SK-CNN, wherein the selective kernel convolutional neural network SK-CNN includes a hidden layer and an output layer, wherein the hidden layer includes an adversarial network DD-GAN model, and the adversarial network DD-GAN model is parallel to the selective kernel convolutional neural network SK-CNN.
[0010] Preferably, in step S3, the soil and meteorological data of the grape field collected in real time are input into a trained deep neural network model, and fertilization and irrigation of the grape field are implemented based on a time iterative algorithm.
[0011] The present invention also provides a wine grape water and fertilizer control device, comprising:
[0012] The acquisition module is used to obtain historical soil and meteorological data of the grape fields;
[0013] A training module for training deep neural network models based on historical soil and meteorological data;
[0014] The control module is used to input the real-time soil and meteorological data of the grape fields into the trained deep neural network model to achieve fertilization and irrigation of the grape fields.
[0015] Preferably, the deep neural network model is a multidimensional parameter prediction model based on a mixture of a generative adversarial network DD-GAN and a selective kernel convolutional neural network SK-CNN, wherein the selective kernel convolutional neural network SK-CNN includes a hidden layer and an output layer, wherein the hidden layer includes an adversarial network DD-GAN model, and the adversarial network DD-GAN model is parallel to the selective kernel convolutional neural network SK-CNN.
[0016] Preferably, the control module inputs the soil and meteorological data of the grape fields collected in real time into the trained deep neural network model, and implements fertilization and irrigation of the grape fields based on a time iterative algorithm.
[0017] The present invention also provides a wine grape water and fertilizer control system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a wine grape water and fertilizer control method when executed by the processor.
[0018] The present invention also provides a storage medium, on which a computer program is stored, and the computer program executes the wine grape water and fertilizer control method when running.
[0019] The present invention trains a deep neural network model based on historical soil and meteorological data; inputs the real-time soil and meteorological data of the grape field into the trained deep neural network model, and implements fertilization and irrigation of the grape field based on a time iteration algorithm. The technical solution of the present invention realizes automatic fertilization and saves labor costs; while reducing the amount of water, the utilization rate of fertilizer is greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0021] Figure 1 The present invention is a flowchart of a method for controlling water and fertilizer for wine grapes according to an embodiment of the present invention. DETAILED DESCRIPTION
[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.
[0023] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0024] Embodiment 1:
[0025] like Figure 1 As shown, an embodiment of the present invention provides a method for controlling water and fertilizer of wine grapes, comprising:
[0026] Step S1, obtaining historical soil and meteorological data of the grape field;
[0027] Step S2: training a deep neural network model based on historical soil and meteorological data;
[0028] Step S3: input the soil and meteorological data of the grape field collected in real time into the trained deep neural network model to achieve fertilization and irrigation of the grape field.
[0029] As an implementation of the embodiment of the present invention, step S1 also includes: cleaning the historical soil and meteorological data. Specifically:
[0030] Screen out garbled data and offside data from historical soil and meteorological data to obtain screening index data;
[0031] Performing vectorization operation on the screening index data to obtain a screening index vector set, and adding a time-space vector to each screening index vector in the screening index vector set to obtain a time-space vector set;
[0032] Performing feature clustering on the space-time vector set to obtain a cluster center data set;
[0033] Fill each empty value data in the screening index data according to the cluster center data set.
[0034] As an implementation method of an embodiment of the present invention, the deep neural network model is a multi-dimensional parameter prediction model based on a mixture of a generative adversarial network DD-GAN and a selective kernel convolutional neural network SK-CNN, the selective kernel convolutional neural network SK-CNN includes a hidden layer and an output layer, the hidden layer includes an adversarial network DD-GAN model, and the adversarial network DD-GAN model is parallel to the selective kernel convolutional neural network SK-CNN.
[0035] Training process: The preprocessed data is divided into a model training set and a model test set in a ratio of 8:2, and then the performance prediction model, i.e., the multidimensional parameter prediction model, is trained with the training samples; the relevant parameters of the multidimensional parameter prediction model are debugged, and the quality of the trained multidimensional parameter prediction model is judged according to the model evaluation index until the optimal deep neural network model is obtained. The determination coefficient R2 and mean square error MSE (mean squared error) can be used to evaluate the training and test results of the model to obtain the optimal deep neural network model.
[0036] Preferably, in step S3, the soil and meteorological data of the grape field collected in real time are input into the trained deep neural network model, and fertilization and irrigation of the grape field are implemented based on the time iteration algorithm. Specifically:
[0037] Input the soil and meteorological data of the grape field at time t into the trained deep neural network model to obtain t+t 1 The soil and meteorological data of the grape fields at all times; the soil and meteorological data of the grape fields include: soil conductivity, soil pH value, soil temperature and humidity, wind direction data, wind speed data, air temperature and humidity, light intensity and rainfall data; t is 7-20 days, t 1 7-20 days.
[0038] According to t+t 1 Keep track of the soil and weather data of the grape fields at all times to control the amount of fertilizer used in a single mixing fertilizer and the amount of water used for a single irrigation.
[0039] Embodiment 2:
[0040] The embodiment of the present invention further provides a wine grape water and fertilizer control device, comprising:
[0041] The acquisition module is used to obtain historical soil and meteorological data of the grape fields;
[0042] A training module for training deep neural network models based on historical soil and meteorological data;
[0043] The control module is used to input the real-time soil and meteorological data of the grape fields into the trained deep neural network model to achieve fertilization and irrigation of the grape fields.
[0044] As an implementation method of an embodiment of the present invention, the deep neural network model is a multi-dimensional parameter prediction model based on a mixture of a generative adversarial network DD-GAN and a selective kernel convolutional neural network SK-CNN, the selective kernel convolutional neural network SK-CNN includes a hidden layer and an output layer, the hidden layer includes an adversarial network DD-GAN model, and the adversarial network DD-GAN model is parallel to the selective kernel convolutional neural network SK-CNN.
[0045] As an implementation method of an embodiment of the present invention, the control module inputs the soil and meteorological data of the grape field collected in real time into the trained deep neural network model, and implements fertilization and irrigation of the grape field based on a time iterative algorithm.
[0046] Embodiment 3:
[0047] An embodiment of the present invention further provides a wine grape water and fertilizer control system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a wine grape water and fertilizer control method when executed by the processor.
[0048] Embodiment 4:
[0049] An embodiment of the present invention further provides a storage medium, on which a computer program is stored, and the computer program executes the water and fertilizer control method for wine grapes when running.
[0050] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.
Claims
1. A method for controlling water and fertilizer for wine grapes, characterized in that: include: Step S1, obtaining historical soil and meteorological data of the grape field; Step S2: training a deep neural network model based on historical soil and meteorological data; Step S3: Input the real-time collected soil and meteorological data of the grape field into the trained deep neural network model to implement fertilization and irrigation of the grape field.
2. The wine grape water and fertilizer control method according to claim 1, characterized in that: The deep neural network model is a multidimensional parameter prediction model based on a mixture of a generative adversarial network DD-GAN and a selective kernel convolutional neural network SK-CNN. The selective kernel convolutional neural network SK-CNN includes a hidden layer and an output layer. The hidden layer includes an adversarial network DD-GAN model. The adversarial network DD-GAN model is parallel to the selective kernel convolutional neural network SK-CNN.
3. The wine grape water and fertilizer control method according to claim 2, characterized in that: In step S3, the soil and meteorological data of the grape fields collected in real time are input into the trained deep neural network model, and fertilization and irrigation of the grape fields are implemented based on a time-iterative algorithm.
4. A water and fertilizer control device for wine grapes, characterized in that: include: Acquisition module, used to obtain historical soil and meteorological data of grape fields; A training module for training deep neural network models based on historical soil and meteorological data; The control module is used to input the real-time soil and meteorological data of the grape fields into the trained deep neural network model to achieve fertilization and irrigation of the grape fields.
5. The wine grape water and fertilizer control device according to claim 4, characterized in that: The deep neural network model is a multidimensional parameter prediction model based on a mixture of a generative adversarial network DD-GAN and a selective kernel convolutional neural network SK-CNN. The selective kernel convolutional neural network SK-CNN includes a hidden layer and an output layer. The hidden layer includes an adversarial network DD-GAN model. The adversarial network DD-GAN model is parallel to the selective kernel convolutional neural network SK-CNN.
6. The wine grape water and fertilizer control device according to claim 5, characterized in that: The control module inputs the real-time soil and meteorological data of the grape fields into the trained deep neural network model, and implements fertilization and irrigation of the grape fields based on a time-iterative algorithm.
7. A wine grape water and fertilizer control system, characterized in that: include: A memory and a processor, wherein the memory stores a computer program executed by the processor, and when the computer program is executed by the processor, the wine grape water and fertilizer control method according to any one of claims 1 to 3 is executed.
8. A storage medium, characterized in that: The storage medium stores a computer program, which, when running, executes the wine grape water and fertilizer control method according to any one of claims 1 to 3.
Citation Information
Cited By
Permanent magnet synchronous motor control method and system based on image processing
CN120074324A