基于树莓派的室内植物病虫害识别系统设计  

Design of house plants diseases and insect pests recognition system based on Raspberry Pi

作　　者：李睿[1] 王莹 王恒 LI Rui;WANG Ying;WANG Heng(School of Electronic and Electrical Engineering,Shanghai University of Engineering Science,Shanghai 201620,China;Intelligent Transportation Division,China Traffic Information Technology Group Co.,Ltd.,Beijing 100032,China)

机构地区：[1]上海工程技术大学电子电气学院,上海201620 [2]中国交通信息科技集团有限公司智慧交通事业部,北京100032

出　　处：《电子设计工程》2022年第8期114-118,共5页Electronic Design Engineering

摘　　要：为了防止由于病虫害导致的室内植物死亡问题,设计了一种基于树莓派的室内植物病虫害识别系统。系统以树莓派3B+为核心,将树莓派3B+与土壤水分传感器和空气温度传感器相连,对土壤中的水分和空气温度进行分析,进而实现基本的灌溉功能。利用摄像头对室内植物进行拍摄,并且将拍摄到的图片用移植到树莓派3B+中训练好的植物病虫害识别模型进行识别,判断是否出现病虫害,并将检测出的植物状态信息显示在屏幕上,同时上传到微信小程序云服务器中,用户可通过微信小程序查看植物病虫害情况。测试实验表明,系统的识别准确率达到91.80%,可靠性较高,具有一定的实用价值。In order to prevent death of plants caused by diseases and insect pests,a house plants diseases and insect pests recognition system based on Raspberry Pi was designed.The system takes the Raspberry Pi 3B+as the core,connects the Raspberry Pi 3B+with the soil moisture sensor and the air temperature sensor,analyzes the moisture in the soil and the temperature in the air,and then realizes the basic irrigation function.In addition,the camera is used to take pictures of indoor plants,and the captured pictures are identified by transplanting to the Raspberry Pi 3B+trained plant pests and diseases recognition model to determine whether pests and diseases are present,and the detected plant status information is displayed on the screen.At the same time,it is uploaded to the WeChat applet cloud server,and users can check the plant diseases and insect pests through the WeChat applet.Test experiments show that the recognition accuracy rate of the system reaches 91.80%,the reliability is high,and it has certain practical value.

关 键 词：树莓派 传感器 病虫害识别 卷积神经网络 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论] TP212[自动化与计算机技术—计算机科学与技术]