Authors: Jameer Shaikh, Dr. Usha B Shete, Dr. A. A. Khan, Dr. R. S. Deshpande
Abstract: Early and accurate detection of fruit diseases is critical for minimizing crop losses and ensuring food security. This study in- troduces a novel automated diagnostic framework that leverages hyperspectral imaging combined with deep convolutional neural networks to detect and classify common diseases affecting apples, including blotch, rot, and scab. By analyzing spectral reflectance patterns from 360 nm to 1000 nm, the proposed method identifies subtle biochemical changes in fruit tissues before visual symp- toms manifest. Extensive laboratory experiments demonstrate that the system achieves an overall classification accuracy of 93.7%, outperforming traditional RGB-based image analysis techniques. Furthermore, field trials conducted in commercial orchards validate the robustness and real-world applicability of the system, revealing a 28% reduction in false positive detections and a 35–40% potential decrease in yield losses through timely intervention. The integration of hyperspectral data with deep learning enables a cost-effective, non-destructive, and scalable solution for precision agriculture, supporting proactive crop management and sustainable farming practices.
