Detection of Cotton Plant Disease Using CNN

Abstract

Cotton production is considered crucial in various parts of the world and determining the diseases well in advance is a vital factor that directly has an effect on the yield. To tackle this issue, a CNN - based approach has been proposed which can detect a diseased plant and the leaf. For detection, the VGG19 artificial neural network has been trained by using google collaboratory. Moreover, unsupervised learning was used with Kaggle cotton plant dataset for training the model followed by validation and testing. Once the training is done, the saved model can easily predict whether the plant or leaf is diseased or not.