Analysis of convolutional neural networks-based approaches in fruit disease detection for smart agriculture applications

Smart agriculture has garnered attention for its potential to optimize resource utilization and enhance crop yield, with video-based fruit disease detection playing a crucial role in mitigating crop losses. This paper offers an overview of current technologies and recent advances in video-based fruit disease detection, with a particular focus on the promising applications of deep learning. Despite notable progress, challenges such as the requirement for large-scale annotated datasets and real-time detection in complex agricultural environments persist. The study contributes significantly by thoroughly examining popular CNN frameworks, including VGG-Net, Res-Net, Inception-Net, and Dense-Net models, through extensive experiments and careful documentation of outcomes. The results provide valuable insights into the efficacy of CNN-based methods for fruit disease detection, emphasizing the study's novelty and paving the way for future research directions in this dynamic field.