Sugarcane Leaf Disease Detection: A Comparative Analysis Using Deep Learning

In this paper, sugarcane leaf diseases were detected through image analysis.This research investigated the effectiveness of four deep learning models, ResNet101, DenseNet161, MobileNetV2, and AlexNet, for sugarcane leaf disease detection.A dataset of sugarcane leaf images was categorized into healthy and various disease classes, and those models were trained and validated on that dataset.The result demonstrated a superior performance which achieved a validation accuracy of 97.41% and a validation loss of 0.0902 by DenseNet161, whereas ResNet101 attained a slightly lower outcome with a validation accuracy of 97.21% and a validation loss of 0.0971.A validation accuracy of 95.82% and a validation loss of 0.1541 were obtained using MobileNetV2.AlexNet obtained the lowest performance with a validation accuracy of 91.43%, and a validation loss of 0.2505, and its performance evaluation was the lowest compared to other models.DenseNet161 and ResNet101 showed high accuracy and low loss for sugarcane leaf disease detection.