Automated Classification of Bean Crop Diseases Using Fine-Tuned ResNet50: A Deep Learning Approach for Enhanced Agricultural Diagnostics

Bean crop diseases substantially lower agricultural productivity since they cause significant losses in production and quality. Early accurate identification of these diseases guarantees food security and helps to lower their load. While considering deep learning techniques to increase the accuracy and efficiency of diagnosis, our work focuses on illness classification in bean leaves. This work uses 8623 photos of bean leaves split into four balanced classes: healthy leaves, angular leaf spot (ALS), bean rust, and unknown class. We employed a modified ResNet50 model for classification. The choice of this architecture was motivated by its established success in picture classification tasks and feature extraction strength. The model was trained and evaluated using this dataset to exhibit an amazing 94% accuracy on the test set, therefore demonstrating its practical use. Since this study provides a reliable, automated method to detect bean crop illnesses, therefore improving crop management methods and maybe reducing the necessity for hand inspections depends significantly on the agricultural sector in this research. This research complements global initiatives aiming at environmental sustainability by supporting innovative agricultural technology, improved food security, and sustainable agriculture itself.