Prediction of Groundnut Leaf Disease Detection and Classification Using Augmented Capsule Neural Network

Groundnut is a significant capital crop grown across over 100 nations worldwide, with India being one of the top producers, including an average yield of 756 kg/ha. The groundnut plant is prone to diseases and viruses which induce damage in the plant, stems, including roots, which reducing output. Early detection, identification, as well as treatment may significantly decrease overall ecological and economic losses. This paper proposed an Augmented Capsule Neural Network (Aug- CapsNet) for classifications and risk stratification estimation in plants. The first approach is designed to diagnose plant diseases using images from plant leaf. Image Enhancement is done by combining the Discrete Fourier Transform (DFT) as well as Thresholding techniques. The Second job is to classify plant leaves. The presented architecture is validated using Plant Village datasets, which includes over 50,000 images representing diseased as well as normal plants. While compared to previous plant disease classification techniques, the Aug-CapsNet model shows significant gains in prediction accuracy. The generated model's experimental outcomes obtained an overall evaluation accuracy of 96.07%, with an F1 score of 95.15%.