Plant Disease Detection and Classification Using Deep Learning Model
B R Pushpa, Adarsh Ashok, Shree Hari A V
Abstract
Organic farming is becoming more common in many developing countries agricultural practices. There are a variety of issues that occur in plant growth due to various environmental factors. Crop diseases can result in a reduction in agricultural productivity and hence identification of crop diseases in the beginning stage will provide great advantages towards the field of agriculture. In the recent years, Identification of leaf disease is done through image processing technique where feature extraction plays important part since using the right features leads to better classification accuracy. Nowadays, there exists lot of machine learning techniques to perform plant disease detection and identification, the advancement of deep learning gained attention due to improved performance and accuracy. So this paper presents a model for detecting and identifying crop leaf disease using CNN based AlexNet model. This proposed model is compared with other CNN model (VGG-16 and Lenet-5) that shows our proposed AlexNet model is more accurate than VGG-16 and Lenet-5. The dataset considered for experimentations are collected from plant village repository with total number of 7070 diseased and healthy leaf images of corn blight, Corn Common Rust, Corn Gray Leaf Spot, Rice Bacterial leaf blight, Rice Brown spot, Rice Leaf smut, Tomato Bacterial spot, Tomato Early blight, Tomato Target Spot and Tomato mosaic virus. The proposed method will successfully identify the crop species with 96.76 % accuracy.