Intelligent greenhouse management system based on computer vision and deep learning

Abstract

The invention relates to the field of intelligent greenhouse management, and particularly discloses an intelligent greenhouse management system based on computer vision and deep learning, and the system comprises a data collection module which obtains environment data and soil data in a greenhouse in real time; the image acquisition module is used for acquiring crop growth images; the core controller preprocesses the data acquired by the data acquisition module and the image acquisition module and uploads the data to the cloud database; the computer vision and deep learning analysis module is used for identifying crop types and pest and disease damage conditions based on the crop images; the water-fertilizer control unit comprises an intelligent control module, a water-fertilizer all-in-one machine and a drip irrigation system; the user interaction module provides remote monitoring and control functions for a user; according to the system and the method, the condition of economic loss caused by the fact that crop diseases and insect pests cannot be correctly treated due to insufficient experience in traditional agriculture is overcome, environmental visualization is realized, a farmer can conveniently monitor the greenhouse in real time, precise irrigation and fertilization are realized, the agricultural production efficiency is improved, and resource waste is reduced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent greenhouse management, and specifically to an intelligent greenhouse management system based on computer vision and deep learning. Background Art

[0002] With the growth of the global population and the increase in agricultural production demands, traditional agriculture faces numerous challenges, such as water resource waste, decline in arable land quality, shortage of agricultural labor, etc. In recent years, intelligent agricultural technologies have developed rapidly. Utilizing advanced technologies such as the Internet of Things, computer vision, and deep learning to improve the automation, precision, and intelligence levels of agricultural production has become an important trend in the industry's development.

[0003] Existing greenhouse management systems mainly focus on the collection and display of environmental data, but they are insufficient in terms of refined adjustment, pest and disease identification, and intelligent decision-making. For example, traditional pest and disease detection methods mainly rely on manual observation, which is not only time-consuming and laborious, but also greatly affected by experience in terms of accuracy; irrigation and fertilization management usually adopt timed or manual control methods, making it difficult to achieve precise fertilization and irrigation, resulting in waste of water and fertilizer resources or restricted growth of crops. Therefore, this application proposes an intelligent greenhouse management system based on computer vision and deep learning to solve the above problems. Summary of the Invention

[0004] In view of the existing problems, the present invention provides an intelligent greenhouse management system based on computer vision and deep learning, which can effectively solve the problems raised in the background art.

[0005] To solve the above problems, the present invention adopts the following technical solutions:

[0006] An intelligent greenhouse management system based on computer vision and deep learning, comprising:

[0007] A data acquisition module for real-time acquisition of environmental data and soil data inside the greenhouse;

[0008] An image acquisition module for acquiring crop growth images;

[0009] A core controller for preprocessing the data collected by the data acquisition module and the image acquisition module and uploading it to the cloud database;

[0010] A computer vision and deep learning analysis module for identifying crop species and pest and disease situations based on crop images;

[0011] A water and fertilizer control unit, which includes an intelligent control module, a water and fertilizer integrated machine, and a drip irrigation system. The intelligent control module controls the water and fertilizer integrated machine and the drip irrigation system according to the analysis results;

[0012] The user interaction module provides remote monitoring and control functions for users.

[0013] As a further solution of the present invention: the data acquisition module includes a light sensor, an atmospheric temperature and humidity sensor, a carbon dioxide concentration sensor, a soil moisture sensor, a pH meter and an EC meter.

[0014] As a further solution of the present invention: the image acquisition module uses a high-resolution camera, which is installed at different positions in the greenhouse to regularly capture crop images. The camera is equipped with an infrared function. The image data collected by the image acquisition module is preliminarily processed by an edge computing device and uploaded to the core controller.

[0015] As a further solution of the present invention: the computer vision and deep learning analysis module uses a convolutional neural network to identify crop species and performs disease analysis based on a crop pest and disease database.

[0016] As a further solution of the present invention: the core controller uses Raspberry Pi to perform data preprocessing, and the core controller receives environmental data and crop images in real time by connecting the data acquisition module and the image acquisition module.

[0017] As a further solution of the present invention: it also includes an early warning module, which automatically sends an alarm message to the user interaction module when an abnormal situation is detected.

[0018] As a further solution of the present invention: the user interaction module includes a Web terminal, a mobile terminal and a PC terminal, and supports remote monitoring and operation.

[0019] As a further solution of the present invention: it also includes a data optimization module, which predicts future irrigation needs through an LSTM neural network based on historical environmental data and crop growth models and generates a dynamic fertilization plan.

[0020] As a further solution of the present invention: the drip irrigation system adopts a pressure-compensated dripper to adapt to terrains with different slopes, and a micro flow sensor is embedded in the drip irrigation pipeline to monitor irrigation uniformity in real time.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention realizes intelligent identification of crop types and pests and diseases through computer vision and deep learning technology, thereby formulating crop cultivation plans, and diagnosing and reporting on crop pest and disease conditions, overcoming the situation in traditional agriculture where economic losses are caused by insufficient experience in correctly handling crop pests and diseases, realizing environmental visualization, facilitating farmers to monitor greenhouses in real time, and combining water-fertilizer integrated machines and drip irrigation technology to achieve precise irrigation and fertilization, improve agricultural production efficiency, reduce resource waste, and promote sustainable agricultural development. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic connection diagram of an intelligent greenhouse management system based on computer vision and deep learning.

[0023] In the figure: 1. Data acquisition module; 11. Light sensor; 12. Atmospheric temperature and humidity sensor; 13. Carbon dioxide concentration sensor; 14. Soil humidity sensor; 15. pH meter; 16. EC meter; 2. Image acquisition module; 3. Core controller; 4. Computer vision and deep learning analysis module; 5. Water and fertilizer control unit; 51. Intelligent control module; 52. Water and fertilizer integrated machine; 53. Drip irrigation system; 6. User interaction module; 7. Early warning module; 8. Data optimization module; 9. Cloud server. Specific implementation mode

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] Combined with Figure 1 This implementation mode is described. This implementation mode provides an intelligent greenhouse management system based on computer vision and deep learning, which includes a data acquisition module, an image acquisition module, a core controller, a computer vision and deep learning analysis module, a water and fertilizer control unit, and a user interaction module. The data acquisition module obtains greenhouse environment data and soil data in real time and uploads them to the core controller. The image acquisition module obtains crop growth images and uploads them to the core controller. The core controller preprocesses the acquired data and uploads it to the cloud database of the cloud server. The computer vision and deep learning analysis module identifies the crop variety and pest and disease situation based on the crop images and provides planting suggestions. The water and fertilizer control unit includes an intelligent control module, a water and fertilizer integrated machine, and a drip irrigation system. The intelligent control module controls the water and fertilizer integrated machine and the drip irrigation system according to the analysis results of the computer vision and deep learning analysis module. The user interaction module provides remote monitoring and control functions for users, and users can view data and perform operations through the mobile terminal and the PC terminal.

[0026] The data acquisition module is deployed in different areas of the greenhouse. The data acquisition module includes light sensors, atmospheric temperature and humidity sensors, carbon dioxide concentration sensors, soil moisture sensors, pH meters and EC meters. The light sensor is installed on the top or side wall of the greenhouse to monitor the light intensity in order to adjust the supplementary light strategy; the atmospheric temperature and humidity sensor is hung in the central area of ​​the greenhouse to detect the air temperature and humidity in real time, providing a basis for environmental control; the carbon dioxide concentration sensor is installed near the greenhouse vent to monitor the CO2 content in the air to ensure the best environment for crop photosynthesis; the soil moisture sensor is buried near the crop root system to measure the moisture content of the soil in real time to provide data support for precise irrigation; the pH meter is installed in different areas of the irrigation water source and soil to detect the acidity and alkalinity of the soil and irrigation water to ensure a suitable growth environment for crops; the EC meter (conductivity sensor) is arranged in the water and fertilizer system pipeline and soil to monitor the fertilizer concentration to adjust the fertilization plan. These sensors continuously collect environmental parameters in the greenhouse and transmit data to the core controller through wireless communication (such as LoRa, Wi-Fi or 4G). Sensor data is used to evaluate the crop growth environment and provide a reasonable basis for water and fertilizer management.

[0027] The image acquisition module uses a high-resolution camera installed at different locations in the greenhouse to regularly capture crop images. The camera is equipped with an infrared function to obtain clear images in low light conditions. The image data is initially processed by the edge computing device, including image noise reduction, edge detection, feature extraction, and format conversion to reduce data redundancy and improve processing efficiency. The processed data is analyzed using a distributed computing strategy and uploaded to the core controller through wireless communication technologies (such as Wi-Fi, 4G, or LoRa) for further deep learning analysis.

[0028] The core controller uses Raspberry Pi as the core unit for data processing. Raspberry Pi receives environmental data and crop images in real time by connecting the data acquisition module and the image acquisition module.

[0029] The core controller works as follows:

[0030] 1) Data reception: Through interfaces such as I2C, SPI or UART, data from sensors such as light, temperature and humidity, CO2 concentration, soil moisture, pH value, EC value, etc. in the greenhouse are obtained, and images taken by the camera are received at the same time.

[0031] 2) Data preprocessing: Convert sensor data into different formats, filter and remove noise, remove abnormal data, and improve data reliability.

[0032] 3) Edge computing: Perform preliminary compression and feature extraction on image data to reduce the amount of data transmitted in the cloud and improve computing efficiency.

[0033] 4) Local caching and uploading: When the network is in good condition, upload the processed data to the cloud database; when the network is interrupted, temporarily store the data and upload it in batches after recovery to ensure data is not lost.

[0034] The preprocessed data is uploaded to the cloud database for further analysis.

[0035] The core controller can also perform local caching to prevent data loss during network interruptions.

[0036] The computer vision and deep learning analysis module analyzes image data based on deep learning algorithms (such as the convolutional neural network CNN). Specifically, the computer vision and deep learning analysis module uses a convolutional neural network to identify crop species and performs disease analysis based on the crop pest and disease database.

[0037] The working principle of the computer vision and deep learning analysis module is as follows:

[0038] 1) Data input: Receive the crop images obtained by the image acquisition module and perform format conversion and preprocessing, such as denoising, color enhancement, etc.

[0039] 2) Feature extraction: Use the convolutional neural network (CNN) model of deep learning to extract image features through multiple convolutional, pooling, and fully connected layers.

[0040] 3) Classification and recognition: Use deep learning models, such as convolutional neural network (CNN) and Transformer architecture, to classify crop species and compare them with healthy crop and disease images in the database.

[0041] The classification process is as follows:

[0042] 4) Feature extraction: Extract the edge, texture, and color features of the image through multiple convolutional operations of CNN;

[0043] 5) Feature classification: Use the fully connected layer (FC) and Softmax classifier to determine the crop species based on the extracted features;

[0044] 6) Disease comparison: Compare the crop image with the pest and disease database and calculate the similarity (using cosine similarity or Euclidean distance);

[0045] 7) Health assessment: Based on the comparison results and environmental parameters, combined with the support vector machine (SVM) or decision tree algorithm, further determine the crop health status.

[0046] 8) Pest and disease detection: Based on the pest and disease database, identify whether the crop has diseases and determine the disease type and its severity.

[0047] 9) Decision Output: The analysis results are stored in the database, and disease warnings or suggestions for adjusting the water and fertilizer plan are provided to the early warning module or the intelligent control module. When the early warning module detects an abnormal situation, it automatically sends an alarm message to the user interaction module.

[0048] The crop species are identified through the model, and the health status of the crops is judged by combining with the pest and disease database. The judgment of this health status is based on the following indicators:

[0049] 1) Leaf Feature Analysis: Detect the leaf color, morphology, and glossiness. If yellowing, spotting, or wilting appears, there may be diseases or nutrient deficiencies.

[0050] 2) Growth Trend Monitoring: Use historical growth data to compare the growth speed and trend of the current crops to judge whether they are in a normal development state.

[0051] 3) Pest and Disease Detection: Combine deep learning algorithms to identify disease spots, signs of pest infestation, etc., and compare with the disease samples in the database.

[0052] 4) Environmental Adaptability Analysis: Combine the environmental parameters (temperature, humidity, light, CO2 concentration, soil humidity, etc.) of the data collection module to judge whether the current environment is suitable for crop growth and infer whether the health of the crops is abnormally caused by environmental factors.

[0053] 5) Water and Fertilizer Supply Evaluation: Compare the EC value and pH data to judge whether the current water and fertilizer supply meets the crop requirements and avoid health problems caused by nutrient deficiency or excess.

[0054] Use large-scale crop growth data to train the model to improve the accuracy of pest and disease detection.

[0055] The training process of the crop growth data training model includes the following steps:

[0056] 1) Data Collection: Use the image acquisition module to obtain a large number of crop images and obtain labeled pest and disease samples from the agricultural database;

[0057] 2) Data Preprocessing: Enhance the image data (such as rotation, scaling, contrast adjustment) to improve the generalization ability of the model;

[0058] 3) Model Training: Use a convolutional neural network (CNN) to train the data set and adjust the weight parameters to improve the classification accuracy of the model;

[0059] 4) Model Validation: Evaluate the accuracy of the model through the test set and optimize the parameters using the cross-validation method.

[0060] 5) Online deployment and update: Deploy the trained model to the cloud or edge computing devices and perform regular updates based on new data to continuously improve the recognition ability.

[0061] Store the calculation analysis results in the database and use them to guide the operation of the intelligent control module.

[0062] The intelligent control module automatically adjusts the greenhouse environment according to the output results of the computer vision and deep learning analysis module.

[0063] Its specific adjustment methods include:

[0064] 1) Water and fertilizer management: Adjust the water and nutrient supply through the water and fertilizer integrated machine according to the crop requirements.

[0065] 2) Temperature and humidity control: Automatically control the fans, heaters or wet curtains based on the temperature and humidity sensor data to improve the stability of the crop growth environment. When the temperature is too high, increase ventilation or atomization for cooling; when the temperature is too low, start the heating device.

[0066] 3) Light optimization: Detect the ambient light intensity through the light sensor and automatically turn on or adjust the supplementary lighting equipment to maintain the optimal light level. At night or on cloudy days, intelligently adjust the intensity of the LED supplementary lights to ensure the photosynthesis requirements of the crops.

[0067] 4) Air quality regulation: Control the CO2 release system based on the data of the carbon dioxide concentration sensor to promote photosynthesis, start the air circulation equipment, optimize the air circulation, and prevent the breeding of germs.

[0068] 5) Remote monitoring and manual intervention: Allow users to view the environmental parameters through the mobile or Web terminal and manually adjust the control strategy. The system provides abnormal warnings and sends notifications to users when the environmental parameters exceed the safe range for timely intervention.

[0069] The specific implementation method of water and fertilizer management is as follows:

[0070] 1) Data monitoring: Real-time obtain data such as soil humidity, EC value, pH value, etc., and analyze the current water and fertilizer requirements of the crops.

[0071] The specific methods are as follows:

[0072] Soil humidity monitoring: Use the soil humidity sensor to measure the current soil moisture content, compare it with the standard crop growth requirements, and judge whether irrigation is needed;

[0073] Electrical conductivity (EC value) analysis: The EC meter measures the electrical conductivity of the soil solution, estimates the concentration of soluble salts in the soil, and evaluates whether the fertilizer content is appropriate;

[0074] pH value measurement: Use a pH meter to monitor the pH value of the soil or irrigation water, compare it with the suitable pH range for the crop, and determine whether it is necessary to adjust the soil environment.

[0075] 2) Data fusion analysis: Combine environmental parameters such as temperature, humidity, light, and CO2 concentration, and use a deep learning model to analyze the water and fertilizer requirements of the crop, and generate optimized irrigation and fertilization strategies.

[0076] 3) Feedback adjustment: Compare historical data with the current state to adjust the water and fertilizer supply strategy to ensure that the crop is in the best growth environment.

[0077] 4) Intelligent decision-making: Calculate the optimal water and fertilizer supply strategy based on deep learning algorithms and historical data.

[0078] 5) Precision fertilization: Control the fertilizer pump to adjust the fertilizer concentration, and use an EC meter to monitor whether the concentration meets the standard.

[0079] 6) Dynamic adjustment: Automatically adjust the irrigation amount and fertilization frequency according to environmental changes (such as temperature, humidity, and light intensity).

[0080] 7) Remote control: Users can remotely adjust the water and fertilizer parameters through the mobile or Web terminal to achieve personalized management.

[0081] 8) Abnormal alarm: When the water and fertilizer supply is abnormal (such as the EC value exceeding the standard or insufficient water), the system will send an alarm and automatically adjust the water and fertilizer supply plan.

[0082] In the water and fertilizer management, combine the data of the EC meter and the pH meter to optimize the fertilization concentration and ensure the balanced nutrition of the crop.

[0083] In the water and fertilizer management, use a water and fertilizer integrated machine to adjust the water and fertilizer supply according to the growth needs of the crop, which specifically includes the following steps:

[0084] 1) Data collection: Use an EC meter, a pH meter, and a soil moisture sensor to detect the soil moisture and nutrient status in real time;

[0085] 2) Intelligent calculation: Based on the sensor data and combined with the crop growth model, calculate the optimal water and fertilizer supply ratio;

[0086] 3) Precision ratio: The water and fertilizer integrated machine adjusts the fertilizer pump and the water flow control valve according to the calculation result to ensure that the mixing ratio of fertilizer and water meets the crop requirements;

[0087] 4) Drip irrigation control: Through the drip irrigation system, it is delivered to the crop roots with a precise flow rate to ensure full absorption of water and fertilizer and reduce waste;

[0088] 5) Dynamic adjustment: According to the change of environmental data, adjust the water and fertilizer ratio in real time to ensure that the crop is in the best growth state;

[0089] 6) Adjust environmental control devices such as fans and heaters in real time through the embedded controller to keep the crop growth in the best state.

[0090] In the water and fertilizer integrated machine, the drip irrigation system adopts pressure compensating drip emitters to adapt to different slope terrains, and the drip irrigation pipeline is embedded with a micro flow sensor to monitor the irrigation uniformity in real time.

[0091] The system also includes a data optimization module, which is connected to the computer vision and deep learning analysis module. Based on the historical environmental data in the greenhouse and the crop growth model, the data optimization module predicts the future irrigation demand through the LSTM neural network and generates a dynamic fertilization plan.

[0092] The user interaction module provides applications for the Web, mobile, and PC sides. Users can remotely monitor the greenhouse situation. The interface provides data visualization functions, including sensor data, crop health status, historical trend analysis, etc. Users can set irrigation parameters, adjust the fertilization plan, and receive abnormal warning notifications through the user interaction module. The user interaction module also supports linkage with other agricultural databases to realize the recommendation of the best planting plan.

[0093] The working principle of the present invention is as follows: A plurality of sensors are used in the greenhouse to sense various environmental parameters of the greenhouse, and the core controller controls the relevant facilities in the greenhouse, thereby changing the biological growth environment in the greenhouse. At the same time, the system uses computer vision and deep learning technologies to construct a convolutional neural network, which can quickly identify plant species, achieve the purpose of formulating specific water and fertilizer supply plans for different crops, and can judge whether the plants have diseases and pests in real time by establishing a disease and pest identification model for relevant plants in advance. Develop corresponding client applications to realize remote monitoring of the greenhouse by farmers. On the client, the recommended irrigation plan obtained by cloud computing will be displayed, and farmers can selectively adopt the system plan or the custom plan. In terms of irrigation facilities, the system selects a water and fertilizer integrated machine and drip irrigation technology to achieve the precision of fertilization amount, and further achieve the purpose of precise fertilization and water quality control, so as to improve the crop quality and yield.

[0094] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0095] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent greenhouse management system based on computer vision and deep learning, characterized in that, Including: A data acquisition module, which is used to obtain the environmental data and soil data in the greenhouse in real time; An image acquisition module, which is used to obtain the crop growth images; A core controller, which is used to preprocess the data collected by the data acquisition module and the image acquisition module and upload it to the cloud database; A computer vision and deep learning analysis module, which identifies the crop species and pest and disease conditions based on the crop images; A water and fertilizer control unit, which includes an intelligent control module, a water and fertilizer integrated machine and a drip irrigation system, and the intelligent control module controls the water and fertilizer integrated machine and the drip irrigation system according to the analysis results; A user interaction module, which provides remote monitoring and control functions for users.

2. The intelligent greenhouse management system based on computer vision and deep learning according to claim 1, wherein, The data acquisition module includes a light sensor, an atmospheric temperature and humidity sensor, a carbon dioxide concentration sensor, a soil humidity sensor, a pH meter and an EC meter.

3. An intelligent greenhouse management system based on computer vision and deep learning according to claim 1, characterized in that, The image acquisition module uses a high-resolution camera, which is installed at different positions in the greenhouse to regularly capture crop images. The camera is equipped with an infrared function. The image data collected by the image acquisition module is preliminarily processed by an edge computing device and uploaded to the core controller.

4. An intelligent greenhouse management system based on computer vision and deep learning according to claim 1 or 3, characterized in that, The computer vision and deep learning analysis module uses a convolutional neural network to identify crop species and conducts disease analysis based on a crop pest and disease database.

5. An intelligent greenhouse management system based on computer vision and deep learning according to claim 1, characterized in that, The core controller uses a Raspberry Pi for data preprocessing. The core controller is connected to the data acquisition module and the image acquisition module to receive environmental data and crop images in real time.

6. The intelligent greenhouse management system based on computer vision and deep learning according to claim 1, wherein It also includes an early warning module, which automatically sends alarm information to the user interaction module when an abnormal situation is detected.

7. An intelligent greenhouse management system based on computer vision and deep learning according to claim 1, characterized in that, The user interaction module includes a Web end, a mobile end and a PC end, and supports remote monitoring and operation.

8. The intelligent greenhouse management system based on computer vision and deep learning according to claim 1, characterized in that, It also includes a data optimization module, which predicts future irrigation requirements based on historical environmental data and a crop growth model through an LSTM neural network and generates a dynamic fertilization plan.

9. An intelligent greenhouse management system based on computer vision and deep learning according to claim 1, characterized in that, The drip irrigation system uses pressure-compensating drippers, which are suitable for different slope terrains, and the drip irrigation pipes are embedded with micro flow sensors to monitor the irrigation uniformity in real time.

Images