Research on cherry shape detection technology based on machine vision

The shape of cherry is one of the most important reference criteria for evaluating the quality of cherry. Whether the stem is intact or not and whether the fruit body is deformed directly affect the economic benefit. At present, the defect sorting of cherries mainly depends on manual work, which has the problems of low efficiency and inconsistent standards. Therefore, a geometric concave-convex method based on computer vision technology was proposed to distinguish normal cherries from deformed ones. Firstly, the original image is preprocessed to extract edge information; secondly, for each edge point, a clockwise point and a counter clockwise one with the same interval are chosen and two vectors are therefore constituted, which are then made cross product to determine whether it's concavity depending on the direction of the normal vector of the cross product. And the total number of concave points is recorded. Then, appropriate thresholds are set to eliminate noise interference and distinguish cherries with deformed shapes. The results showed that the concave-convex method could detect 100% deformed and normal cherry fruits.