Complete surface information of an individual fruit is essential for automatic detection based on machine vision technology. In this research, two mirrors were placed on the both sides of the conveyor. Individual apples were transmitted on the conveyor in three ways: 1) the calyx and stem were horizontal 2) the calyx and stem were vertical 3) postures of apples were at random. The angel of the mirrors and the speed of the conveyor are both adjustable. All the experiments are conducted in on-line environment. We try to find a suitable method in which complete surface information of apples could be acquired efficiently and practically. The experimental results show that it works best when the calyx and stem are horizontal.
展开▼
