1996 Volume 58 Issue 4 Pages 35-43
Image recognition experiments of crisp head vegetables were done to develop a selective harvesting robot. Recognition of 2-D image of crisp head lettuces was first tried by using a TV camera according to the difference of the shadow pattern. It was possible to check up the crisp heads above a proposed diameter and to measure their rough sites. However, in some conditions, the recognized site and the size were not accurate. For this reason, the manufacture of 3-D vision sensor was tried to directly analyze the shape of crops and preparatory experiments were done. The sensor could measure the 3-D shape of lettuces in 4 seconds.