Maize and Peach Leaf Disease Detection Using Image Processing and Machine Learning

The reason for detecting plant diseases in agriculture is that the early detection prevents the loss of crops on a large scale. Working with wide ranges, the time taken by plant diseases to be identified manually is too long, and it might be less accurate. This paper introduces an automatic methodology that can identify peach and maize leaf disease by the use of CNN. In our approach, the steps involved are pre-processing, segmentation, feature extraction, classification, and image acquisition. This paper aims at automatically identifying plant illnesses through crop disease detection and management. We will investigate the effectiveness of the CNN architectures based on accuracy and inference time for ResNet, MobileNet, and EfficientNet compared to traditional machine learning methods such as SVM and Random Forest. The new architecture is proven to be a promising candidate in practice due to its high accuracy at 97.83% with respect to the existing methods.