Applying Deep Convolutional Neural Network for Classification of Black Gram Plant Leaf Disease

Black gram plant leaf disease can cause significant crop yield reductions, resulting in economic losses and affecting food security as well as impacting agricultural sustainability. Early detection of black gram plant leaf disease is essential for reducing losses and maintaining crop quality. In recent years, advanced approaches in deep learning and computer vision algorithms have become promising tools for automated plant disease identification. In this study, we have applied a convolutional neural network (CNN) based architecture called BPLD-10 to deal with the problem of detecting black gram plant leaf disease. Utilising the Black gram Plant Leaf Disease (BPLD) dataset, which contains 1000 high resolution images classified into five classes, including a healthy leave and four disease types, this paper focuses on enhancing the model overall accuracy and effectiveness. We have initiated the process by applying normalisation image preprocessing techniques for optimising disease detection as well as data augmentation methods for dataset balance and model robustness. The proposed CNN model named BPLD-10 beat existing state-of-the-art models, obtaining a stunning test accuracy of 98.66%. This research helps to early disease diagnosis by providing an efficient and reliable approach to black gram plant leaf disease detection, enabling farmers to take early steps and enhance crop yields.