Establishment of an Automatic Crop Classification and Disease Detection System: Applied to Apple and Tomato Disease Detection.

Agriculture in Taiwan is facing problems such as population aging and labor shortage. Therefore, effectively managing the crops in the planting process is often impossible, easily exposing them to disease threats and leading to heavy losses. This study builds a crop disease detection system and applies it to apple scab, tomato early blight, late blight, and leaf mold. The system first collects apple and tomato crop photos and crop disease data from the Plantvillage dataset on the Kaggle platform. Secondly, according to the crop dataset, the crop type recognition and leaf detection model is trained to identify crop types and detect plant leaves. Finally, the detected leaves are inputted into the disease recognition model for disease detection, and the plants infected with the disease are found. The crop detection system intends to help farmers improve the efficiency of monitoring crop disease through aerial cameras so that farmers can treat plants early and avoid agricultural impacts caused by crop disease. [ABSTRACT FROM AUTHOR]