Research on agricultural disease detection technology based on artificial intelligence

Abstract. Pests and disease are two major factors that reduce the yield of agricultural products per acre of arable land. A report from the Food and Agriculture Organization of the United Nations suggests that agricultural disease and pests cause a loss of more than 1/3 of agricultural production every year. However, farmers cannot check their crops all day without a break. Artificial Intelligence (AI) is developing rapidly this year, it performs well in image recognition. Thus, it is suitable to integrate AI in agriculture. This paper shows the outcome of the accuracy of the outcome when a model is trained using the ResNet-50 model and by some crop pathogens datasets downloaded from Kaggle and using a platform of mmpretrain. It also discusses the pros, cons, and areas for improvement when using AI to detect agricultural disease. This research aims to enhance crop yields and agricultural stability by helping integrate the model in agricultural production.