This thesis explores the development of an automated vision-based quality control system for roasted cashew processing. Traditional quality inspection methods rely heavily on manual labor, which is time-consuming, inconsistent, and prone to human error. The proposed system leverages image processing and machine learning techniques to accurately detect and classify roasted cashews into three quality categories: burnt, unroasted, and good. Using a simple hardware setup—an HD webcam and a light—the system ensures cost-effectiveness and accessibility for small to medium-sized enterprises (SMEs). The study evaluates the system's performance under various lighting conditions and assesses its adaptability to realworld challenges. The results demonstrate the potential of the system to enhance operational efficiency and product consistency, making it a viable alternative to traditional inspection methods.