No abstract

This model arrives at the solution of resolving plant disease recognition by constructing a convolution neural network (CNN) and its pre-trained architectures which gave higher efficiency of 99%. Though there are many models such as SVM, Res-Net has been used to classify the same dataset of paddy leaf diseases, CNN gave the better accuracy. This model helps farmers to classify a large sample of data with accurate classification there by making it cost efficient and saves the loss that could possibly occur when the disease is left unrecognized. Res-Net is easy to train because it is faster and has a smaller number of parameters compared to VGG-16. The dataset has three main diseases, namely bacterial leaf blight, brown spot, and blast. The data's are equally divided among different classes by performing augmentation. Since there is equal number of augmented images, enough features will be extracted from the dataset. The cause of the disease may be because of any source such as meteorological factor as well as pests but detecting at earlier stages by classifying correctly would avoid the destruction of crops. The more the images for every class more accurate the result will be.

No abstract

No abstract