Cotton Plants Disease Identification using Image Segmentation and Machine Learning

Abstract

In India s agricultural economy, cotton is one of the most crucial cash crops. It plays a vital role in the livelihoods of over 6 million farmers, contributing significantly to their income and sustenance. Despite its importance in India s agricultural economy, cotton newlinecultivation faces a significant disadvantage in the form of high susceptibility to pests newlineand diseases. Research shows that approximately 30-35% of the cotton crop yield is newlineprone to contamination, highlighting the need for effective pest and disease management newlinestrategies to ensure optimal crop yield and quality. The timely identification of newlinediseases in cotton cultivation is crucial since detecting diseases at an advanced stage newlineoften results in crop failure. However, recent computer vision and machine learning advancements present a promising solution that can significantly impact the agricultural industry. By leveraging these technologies, farmers can identify and address crop diseases early on, enabling them to take corrective measures to prevent crop failure and increase yield. The symptoms of most diseases affecting cotton plants are typically visible on the leaves, flowers, and fruits. In a recent research study, multiple studies were reviewed, which utilized image pre-processing, segmentation, and machine learning techniques to identify and categorize various diseases to distinguish peculiar cotton diseases. newlineThe study s findings suggest that advanced machine learning approaches outperform newlinecommonly used image processing methods in terms of accuracy and other proper newlinemeasures. The study aimed to identify and classify disorders affecting cotton leaves, newlineproviding a potential solution for early disease identification in cotton cultivation. The results reveal that the model distinguishes healthy and unhealthy cotton leaves. CNN with a convolution layer achieves a higher accuracy (96.25%) than ANN and SVM. newline