Paper Key : IRJ************348
Author: Mirza Mohammad Abdullah Wasim Ahmed Baig
Date Published: 25 Oct 2023
Abstract
Small farmers are still dealing with diseases that ruin their crops. This is dangerous because it can destroy their livelihoods and impact the amount of food available to people globally. Theres hope though. The number of smartphones on the market is increasing, and with more people having access to this technology theres an opportunity to use it for agriculture research. Especially since computer vision models have shown potential in solving agricultural problems. So, that's what we did in this study. We examined how well bespoke CNNs (Convolutional Neural Networks) can detect diseases by using pre-trained models like Inception Model V3, Resnet Model, VGG16 model and VGG19 Model on around 20,600 photos of tomato, potato, and pepperbell leaves. With this we were able to create a model that has detection rates of 92%, 94%, 97% and 96%. Its important for us to put these into practice as soon as possible so small farmers and sustainable agriculture can benefit from them. With food security being at risk now more than ever due to the ever changing landscape of agriculture, our work is essential in increasing crop output, reducing economic losses and eventually guaranteeing food security. Keywords: Analysis, Detection, CNN, Models, Disease.