Control of Sealing Robot Using Seam Position Detection Method Based on Integration of Range and Gray-Scale Images.

Abstract

A seam position detection method is developed for a sealing robot. At the start of sealing, a six-degree-of-freedom seam position is detected by using integration of an optical profile and grayscale images. After the sealing has started, a three-degree-of-freedom seam position can be detected by using only the optical profile image. However, a deviation between the actual seam position and the desired position taught beforehand exists. In addition, the time required for image processing is considerably long compared to the sampling period in a robot control system. Thus, a holder between the vision sensor signal and the robot control signal is necessary to counteract these deviations. A control method adopting a 2nd-order holder utilizing previous signals is developed to correct the deviations of seam positions determined by image sensing and, at the same time, it corrects the deviations without time delay due to its use of forward detection by image sensing.