Crop Health Monitoring: YOLACT Deep Learning Model and Histogram Equalization for Enhanced Leaf Blight Detection

Leaf blight, a widespread plant disease caused by diverse pathogens (fungi, bacteria, viruses), significantly reduces agricultural productivity through leaf browning, wilting, and decay. Early detection is critical to prevent yield loss, but manual methods are labour-intensive and error prone. This study proposes a novel deep learning approach for automated leaf blight detection and identification. We leverage the YOLACT model, demonstrating high precision in recognizing various plant diseases, including leaf blight. To address challenges like light variations, shadows, and overlapping leaves, we incorporate histogram equalization as a pre-processing step to enhance image contrast and detail. This combined approach, utilizing YOLACT and histogram equalization on camera-captured leaf images, achieves a detection accuracy of 98.94%. This innovative methodology offers an efficient solution for leaf blight detection, contributing to effective disease management and improved agricultural productivity.