基于高精度传感器和卷积神经网络的智能灌溉系统研究  

Research on intelligent irrigation system based on high-precision sensors and convolutional neural networks

作　　者：徐世周 鲁宸硕 张梦洁 程霄霄 钟一鸣 Xu Shizhou;Lu Chenshuo;Zhang Mengjie;Cheng Xiaoxiao;Zhong Yiming(College of Electronic and Electrical Engineering,Henan Normal University,Xinxiang 453007,China)

机构地区：[1]河南师范大学电子与电气工程学院,河南新乡453007

出　　处：《河南师范大学学报（自然科学版）》2024年第4期52-56,93,I0007,I0008,共8页Journal of Henan Normal University(Natural Science Edition)

基　　金：国家自然科学基金(11904085);河南省科技攻关项目(202102210299).

摘　　要：为了实现智能精准灌溉,以STM32为主控板,将基于卷积神经网络图像识别技术应用在灌溉系统中.结合EC-5土壤水分和温度传感器来检测土壤的湿度和实时温度,利用图像识别技术对土壤不同的干旱情况下植物叶片状态的识别.通过结合农作物叶片状态和土壤的湿度情况进行精准灌溉.在Keil5中完成各个功能模块的软件设计,通过主控板将采集到的数据与设定值做对比,实现对植物的精准灌溉.同时,用户可以在手机端上实时观察土壤湿度和植物的状态,也可以通过手机端进行远程操作.经测试,本系统可以实现对植物的精准灌溉.In order to realize intelligent and precise irrigation,image recognition technology based on convolutional neural network was applied in irrigation system with STM32 as the main control board.Combined with EC-5 soil moisture and temperature sensors to detect soil moisture and real-time temperature,image recognition technology is used to identify plant leaf status under different soil drought conditions.Precision irrigation is achieved by combining crop leaf state and soil moisture.The software design of each functional module was completed in Keil5,and the collected data was compared with the values set through the main control board to realize precise irrigation of plants.At the same time,users can observe the soil moisture and plant status in real time on the mobile phone,and can also operate remotely through the mobile phone.The test shows that this system can realize the precise irrigation of plants.

关 键 词：STM32主控板 卷积神经网络 EC-5土壤水分传感器 植物状态 灌溉 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]