2004 Volume 66 Issue 3 Pages 59-68
The objective of this study is to develop an motion recognition system which can think an intelligent support autonomously by tracking worker's behavior in agricultural field. The effectiveness of the worker tracking method by interframe difference and that of the motion expressing method by optical flow were investigated on some kinds of typical worker's behavior in this paper. The tuning of some parameters of both methods, the upgrading of frame rate and the reduction of calculation cost were essential not only to track workers but also to calculate the optical flow in real time.