Identification and Prediction of Sugarcane Diseases using Convolution Neural Network

Sugarcane is a major global crop for sugar and bioethanol, produced mostly in tropical and subtropical countries, with Brazil and India leading producers. These countries provide 80% of the world's sugar and 40% of the bioethanol output. A multitude of environmental variables can lead sugarcane to become vulnerable to a wide range of diseases, such as Bacterial Blight, Mosaic, Red Rot, Rust, and Yellow. For diagnosis of sugarcane diseases, an accurate and reliable automated disease detection system is required. In this study, we developed convolution neural network model EfficientNetB0 with data augmentation, compared its performance, and assessed its efficacy in recognizing illnesses that affect sugarcane plants using leaf photos. The EfficientNetB0 model has precision, recall, and F1 score values of 0.9812, 0.9821, and 0.9816. With a 98.12% accuracy rate, this model is useful tool that growers may use to keep the sugarcane diseases from interfering with the sugarcane harvesting process.