Eggs Detection and Classification Using YOLOV5

Abstract

As poultry product quality increases, so does the need for effective poultry egg sorting due to the limitations of manual processes. Modern farms now utilize automated systems to enhance productivity and accuracy, as well as improve animal welfare. An intelligent egg detection and classification system based on deep learning is proposed in this study. A dataset consisting of white and brown chicken eggs was collected and annotated alongside multiple variants of YOLOv5 to train and evaluate them. Various metrics including precision, recall, F1 score, mAP, and computational time were measured to determine how effective each model was. Results showed that the YOLOv5n model outperformed the rest with an F1 score of 0.98, along with achieving excellent detection accuracy and low computational requirements, thus showing suitability for real time applications. The work done in this paper demonstrated the possibilities given by computer vision in automating egg sorting and laid the groundwork for applying such systems into fully autonomous poultry farms.