An Automatic Fine-Tuned InceptionV3 Model for Groundnut Leaf Disease Multiclass Classification

There are various diseases which attack the groundnut crops resulting in the worst quality of nut production. To create an automatic method for spotting them early, in the proposed work, a deep learning-based InceptionV3 model has been developed. This will help stop big yield losses due to the disease's effect on the crop. This study looks at using the InceptionV3 convolutional neural network (CNN) for detecting diseases in groundnut leaves. It compares how well Stochastic Gradient Descent (SGD) and Adam optimizers do the prediction at different numbers of epochs. The study uses a large collection of data that includes five diseased and healthy classes of groundnut leaves. This model shows excellent findings for groundnut leaf disease detection, which gets 97.30% accuracy with SGD and 96.56% correct prediction using an Adam optimizer. In this work, the results have been depicted with varying numbers of epochs. The proposed InceptionV3 is a popular deep-learning design for feature extraction, that helps the model spot small patterns and shows different types of diseases. This study shows the outcomes in terms of metrics such as accuracy, recall, precision, and F1 score. It helps to better understand how well an Inception V3model works for different disease types classification. The results show that when the Inception V3model is implemented with SGD, it performs well for groundnut leaf disease categorization. The results show its potential to be used for real-world tasks related to agriculture where the optimal classification is the matter of concern. Lastly, the results show that Inception V3is good for finding diseases in groundnut leaves on the basis of various parameters. The work also shows, how important it is to choose a correct optimizer method for a particular problem statement. The accuracy rates of 97.30% with SGD and 96.56% with Adam show how good the model is for real use in farms, it can also help farmers in classifying the diseases early on so they can manage them better without the production loss.