An Intelligent Optimization Approach for Mango Disease Detection

Abstract

Fruit diseases pose a significant challenge to agricultural productivity and fruit quality, resulting in substantial economic losses worldwide. Therefore, the timely and accurate detection of fruit diseases is essential for improving crop management, minimizing production losses, and promoting sustainable agricultural practices. This research proposes an efficient mango fruit disease detection framework using a Capsule Network integrated with the Chaotic Puma Optimization Algorithm. Initially, the collected mango image dataset undergoes preprocessing techniques to improve image quality and enhance disease-related features. The preprocessed images are then provided to the Capsule Network, which consists of Primary Capsule and Deep Capsule layers for effective feature extraction and preservation of spatial relationships among disease characteristics. To further enhance the model performance, the Chaotic Puma Optimization Algorithm is employed for optimal parameter tuning, improving convergence and minimizing classification errors. Finally, a Softmax classifier is utilized to classify mango images into healthy and diseased categories based on probability scores. The effectiveness of the proposed framework was assessed using standard classification performance metrics. Experimental results demonstrate that the proposed model achieves a high detection accuracy of 98.8%, proving its effectiveness and reliability for automated mango fruit disease detection applications.