A Hybrid Method of Level Set and Extreme Point Tracking Algorithm for Fruit Diameter Measurement

摘要

Recording changes in the fruit size can he used to evaluate the fruit growth situation. To realize the non-contact and accurate measurement of fruit size, a fruit diameter measurement method is proposed. Since it happens that fruits in sampled images are often obscured by steins and leaves, which affects the efficiency of segmentation and extraction. Thus in this method, we first use the extreme point tracking algorithm to acquire the initial contour curve for level set method. Finally, through the level set method, effective segmentation of melon fruit is achieved. Experiments have proved this method has advantages of both non-contact and high-precision.