Deep Learning-Based Detection of Plant Diseases

B. Rajasekharam¹, Voruganti Anuradha², Madem Satya Nooka Vijay Kumar³,
Saka Mary Jasmine⁴, Kashv Yadav⁵, Maradana Bhanu Prakash⁶
¹Assistant Professor, ²³⁴⁵⁶Students
¹²³⁴⁵⁶Department of Computer Science and Engineering
Visakha Institute of Engineering and Technology
Narava, Visakapatnam
Andhra Pradesh, India

Abstract:Plant diseases pose a significant threat to agricultural productivity and global food security, making early and reliable detection essential. This paper presents a deep learning-based approach for automated plant disease detection using a fine-tuned ResNet50 model. The system classifies plant leaf images into 38 categories, covering both healthy and diseased conditions, and achieves a validation accuracy of 99.13%, demonstrating strong generalization capability. To enhance model transparency, Grad-CAM is integrated to visually highlight disease-affectedregions, enabling users to understand the reasoning  behind predictions. The proposed model is deployed as a web-based application that provides real-timepredictions along with confidence scores, visual  explanations, condition analysis, and treatment recommendations. In addition, the system incorporates multilingual support and crop-specific information to improve accessibility and usability. The results indicate that the proposed approach is not only highly accurate but also practical and interpretable, making it suitable for real world agricultural applications.Keywords: Deep Learning, Grad-CAM, Precision Agriculture, Transfer Learning, Plant Disease
Detection, ResNet50.