Chili Spur Pepper Detection Model Developing for Sorting Machine

Abstract

Faculty of Science at Sriracha, Kasetsart University Siracha Campus, located in an industrial and agricultural area, aims to advance the agricultural industry through the application of computer technology. This research focuses on improving the sorting process for chili spur peppers by utilizing deep learning techniques such as YOLO to develop a model that accurately analyzes and sorts of peppers by color. The model is intended to enhance sorting processes in production lines and packaging within the agricultural industry. Researchers applied the YOLO technique, collecting 294 images of chili spur peppers and evaluating the model's performance using metrics such as Precision, Recall, and Mean Average Precision (mAP). The results demonstrated a high accuracy, with an average Precision of 99.6%. The study also evaluated the model's practical application, considering factors like accuracy, speed, and economic efficiency, showing that it could be effectively implemented in chili sorting machines. However, further refinement is needed to improve accuracy in certain environmental conditions. Additionally, the research provides insights for the further development and application of image detection technology for sorting other agricultural products, aiming to increase efficiency and reduce waste in agricultural production processes.