Increasing grain production is essential to those areas where food is scarce. Increasing grain production by controlling crop diseases in time should be effective. To construct a prediction model for plant diseases and to build a real-time application for crop diseases in the future, a deep learning-based image detection architecture with a custom backbone was proposed for detecting plant diseases. In order to get a good amount of crop we need to detect the disease at the earliest. Basically crop disease diagnosis depends on the different characteristics like color, shape, texture etc. Here the person can capture the images of the crop and then the image will be sent to the trained YOLO model. The model analyzes the image and detects crop disease like Blast, False smut, Sheath Blight .