In order to help people solve crop health problems and improve the efficiency and quality of agricultural production,this research designed and implemented a multi-module image detection system for agricultural field based on ResNet50 and YOLOv5 deep learning algorithms,and used PyQt5 technology for visualization. Through data enhancement,learning rate optimiza-tion,hyperparameter adjustment and transfer learning operations,the ResNet50 and YOLOv5 models have achieved accurate detec-tion of crop pests,diseases and health;the verification results show that the system has achieved a good recognition level in disease identification,water shortage identification,trace element deficiency identification,toxic plant identification and weed detection mod-ules,which proves the feasibility and practicality of the system.
关键词
深度学习/ResNet50/YOLOv5/病害识别/杂草检测
Key words
deep learning/ResNet50/YOLOv5/disease identification/weed detection
维普
万方数据