Revolutionizing Plant Disease Detection with CNN and Deep Learning

Recent advancements in plant leaf disease detection using image analysis have sparked a reevaluation of farming practices, especially in agriculture-dependent countries like Bangladesh. Automatic detection systems provide timely identification of roots, stems, leaves, and fruits, allowing farmers to capture and diagnose plant diseases through images easily. This approach saves time and reduces costs while enabling remote monitoring and support for farmers. Our paper introduces a CNN-based technology for plant disease detection and assesses its performance using various algorithms, including Alexnet, Googlenet, Resnet50, VGG16, VGG19, Darknet53, Shufflenet, Squzzenet, Mobilenetv2, Inceptionv3, Inceptionresnetv2, Efficientnetb0, Desnet201, Xception, Nasnetlarge, Yolov7, and Nasnetmobile. Among these, Yolov7 exhibited exceptional validation accuracy of 96.1%, showcasing its effectiveness in disease localization and identification. This study highlights the necessity of deep learning optimizers in improving picture classification outcomes, particularly for plant disease diagnosis, and recommends improvements in deep learning architectures and optimization techniques.