| In recent years,with the rapid development of China’s economy,people’s demand for green,healthy,high-quality crops is growing,therefore,the requirement of modern agricultural technology is higher and higher.In the process of agricultural production,many natural factors affect the normal growth of crops.At the same time,because of the wide variety of plants in modern agricultural planting areas,managers hard to grasp the growth of crops in real time and effectively,such as mature cycle,pests and diseases,resulting in the inevitable trouble.Therefore,with the rapid development of Internet of things technology,the application of Internet of things technology,computer vision technology and intelligent monitoring technology in agricultural production plays a vital role in the intelligent supervision of modern agriculture.In this paper,a plant vision monitoring system based on embedded Linux is designed,which uses the LoRa wireless technology to realize the real-time monitoring of the parameters of the growing environment of crops,and uses the camera in the monitoring equipment to capture the images of crops,combining with the NBC leaf recognition classifier based on CSA designed in this paper,it can effectively identify the crop species,in addition,it can be adjusted and controlled in real time by the visual monitoring system to ensure that it is not disturbed by the complex environmental factors,finally,the remote monitoring and management of greenhouse crops will be realized.The work done in this article is as follows:(1)The monitoring and control terminal is designed and completed.LoRa communication technique is proposed and used to build the whole visual monitoring system.In view of the low power consumption and wide coverage area of LoRa,the terminal can communicate with the server in real-time and effectively.At the same time,the hardware of each part of the system is designed,including core microcontrol module and the LoRa wireless module.(2)The software is designed for the field need of visual monitoring system.Mainly including embedded software and related design,built a storage greenhouse environment parameters SQLite database;To view the data of greenhouse environment,capture the image of crops,record the scene of monitoring,query the historical data and carry out intelligent control on the monitor interface designed on the basis of Qt Development Platform,it realizes the data interaction between the greenhouse environment parameters and the server and the control of the core microcontroller on the greenhouse site(3)In view of the variety of crops in the daily planting process and other factors will affect the quality of greenhouse cultivation,real-time detection efficiency and other iss.To make the following measures: the camera capture to crop the image in a centralized controller as preprocessing,and then combined in this paper,the design of naive bayesian classifier based on clonal selection algorithm for crop leaf sample training,which can effectively identify the corresponding types of crops,greatly improve the recognition accuracy and recognition;By observing the crop images recognized by the classifier to identify whether the plant has diseases and insect pests,it can improve the work efficiency of the planting personnel,and the final recognition results are placed on the interface of the upper computer for viewing.(4)The installation and test of the system have been completed.The experimental results show that the data transmitted by LoRa is stable and of good quality,and can be displayed in real time and accurately on the monitoring interface of the host computer,the measurement error of the sensor is small,and the recognition accuracy of the blade classifier is high,meet the identification requirements.The system can meet the basic requirements of visual monitoring for greenhouse crops. |