A top-mounted mirror-based imaging system for rapid multi-view quality inspection of fruits

Computer vision is a powerful tool for external quality assessment of fruits, and rapid acquisition of surface images is essential for practical applications. In this study, we propose a rapid multi-view fruit surface imaging system that uses a single camera combined with a top-mounted mirror-based configuration. By properly arranging the mirrors, both the top and lateral surfaces of the object are reflected and captured within a single shot. The performance of the system was evaluated using spherical and cylindrical samples as well as real fruits, including pomelos and mangoes. Experimental results demonstrated that more than 80% of the fruit surface, including the entire top region and most of the lateral surface, could be captured in a single image. Defects in challenging areas such as the equator, stem, and calyx were clearly visible. In contrast to existing solutions using multiple cameras or a single camera with rotating fruits, the proposed system reduces hardware costs and data volume while significantly increasing acquisition speed and overall efficiency. Moreover, positioning the mirrors above the fruit enables easy integration into conveyor-based automated inspection systems commonly used in industry. These findings highlight the strong potential of the proposed system for practical fruit inspection tasks such as volume estimation, shape analysis, color uniformity assessment, and surface defect detection.