2013 年 64 巻 2E 号 p. 251-257
In order to prevent human errors in the manufacturing industries, in this paper, a detection method of human error using image analysis is considered using screw tightening work as the target work. The results show that when the worker grasps the driver, the hand has a characteristic where the fingers from second to fifth fingers are parallel and also the high part and low part of the brightness value occur alternatively. It is discovered that through use of this characteristic for image analysis, the position of the hand while grasping the screwdriver is accurately estimated and screw tightening work is analyzed.