Smart Machine Learning Based Cotton Leaf Disease Detection System using Raspberry Pi Manek Asha S1, Vineeta2*, Mishra Pranay3 1Department of Computer Science Engineering, RV Institute of Technology & Management, Bengaluru, India 2Department of Computer Science Engineering, AMC Engineering College, Bengaluru, India 3CoP-DE IT Outsourcing, Hitachi Vantara, Bengaluru, India *vini.upmanyu@gmail.com
Online published on 30 November, 2021. Abstract Indian farmers cultivate cotton on a large scale as it is among vital cash crops. The major problem which leads to reduced production is the attack of disease on plants. Most of the diseases are seen on leaves, flowers and fruits in cotton plant. Early diagnosis is important for identifying cotton diseases. Monitoring the health of cotton leaves is very difficult by naked eyes. A smart approach using machine learning is proposed to resolve this problem which can evaluate the images of the leaves of the plant and spot the disease. The proposed system uses Raspberry Pi and image processing techniques for diagnosing diseases in cotton leaves. Classification is done by choosing suitable features such as color, texture of images etc. by applying different preprocessing techniques such as image filtering, removal of background and image enhancement. Colour based segmentation is done to get the diseased part from the leaves and these segmented images are used in feature extraction. Classification algorithms such as the K-Nearest Neighbor (K-NN), Support Vector Machine (SVM) and Artificial Neural Network (ANN) are implemented to get the accurate results. Top Keywords Artificial Neural Network (ANN), Diseased Cotton leaf, Image pre- processing, K-Means Clustering (k-Means), K-Nearest Neighbor (KNN), Support Vector Machine (SVM). Top |