Enhancing Potato Crop Health: Accurate Disease Classification through Deep Learning and Image Processing

Abstract

Potato plants are primarily affected by fungus, resulting in early and late blight diseases and reducing the production rate of crops. Real-time identification and management of diseases could help farmers enhance production and reduce financial losses. This study introduces a time-efficient algorithm using image processing to accurately classify diseases caused by Alternaria Solani and Phytophthora Infestans in potatoes. Our methodology comprises three steps: pre-processing (grayscale conversion, enhancement), image segmentation (soft clustering, morphological dilation, and flood fill operation), and classification (AlexNet). This framework is tested over the three classes of the PlantVillage dataset of the potato crop. Experimental results demonstrated satisfactory results with 97.57% accuracy in 30 minutes and 58 seconds.