Wheat stripe rust is a major agricultural biohazard of epidemic nature affecting the safety of wheat production caused by the stripe rust fungus, characterized by fast spreading speed, strong outbreaks, heavy losses, and great difficulty in prevention and control, which often causes devastating damage to wheat production in epidemic years. Since the disease can be controlled at an early stage by spraying and other measures, how to effectively predict the occurrence of wheat stripe rust at an early stage has been highly valued by agricultural plant protection departments. To alleviate the above problems, this project explores a deep learning-based prediction system for wheat stripe rust.
Based on deep learning, this project obtains the prediction of wheat stripe rust disease prevalence by acquiring data on meteorological data of previous years, wheat variety planting, stripe rust fungus source and other key factors of prevalence, combining with multispectral remote sensing images from UAVs and RGB data, and utilizing an artificial intelligence prediction model based on the improved BP neural network algorithm as well as the Yolov8 target detection algorithm to obtain the prediction of wheat stripe rust disease prevalence. Based on this prediction value, the prevention of wheat stripe rust disease and other work is done in order to achieve the purpose of reducing the loss of wheat yield and bring indirect economic benefits.
Detail
Shen Yaxuan
Cui Jiajun
Hu Zhiwei
Chen Jiawang
Pan Jiayi
Xu Yifei