YOLO8 Based Smart Detection of Contaminants in Hyacinth Bean Grains

Precise quality assessment of good grains is crucial for effective crop management and improving yield. Traditional methods often face issues like subjectivity and inconsistency, which can be overcome with automation. This study explores using the YOLOv8 object detection algorithm to automate the grading process for hyacinth beans grains. Work was carried out on our dataset images labelled for various bean qualities including high-quality beans, defective ones (Weevilled), and contaminants like husks, stick, other mixed grain, stone, and insects. The YOLOv8 model was trained to identify and classify these beans and contaminants accurately. The automated system significantly reduces manual labor and minimizes errors, offering a reliable and efficient solution for quality assessment. By automating the quality assessment of hyacinth beans, this research supports precision agriculture, leading to more sustainable food production and economic benefits.