ABSTRACT
Summary
Mobile robots are sophisticated mechatronic systems. Sensing, communication, and control technologies of mechatronic systems are applicable in multi-robot cooperative tasks. In this chapter, a fast computer vision algorithm is presented for a multi-robot cooperative control system. This algorithm identifies the current poses (positions and orientations) of a robot and the manipulated object (a rectangle box) from a color image, in real time. Two main challenges are faced in the multi-robot task considered here. The first one concerns the response speed of the vision subsystem. Ordinary computer vision algorithms available in the literature usually are rather slow and unacceptable in the present real-time system. The second challenge comes from uneven lighting, which makes it very difficult for the vision subsystem to trace the same color blob in different positions. A fast computer vision algorithm is presented to cope with these challenges. First, an image in the RGB (red—green—blue) color space is converted into the HSI (hue—saturation—intensity) color space. Then the saturation and the intensity components of the image are removed, and only the hue component is retained. This operation effectively removes the disturbances caused by uneven lighting conditions in the environment when the robot and the object move from one position to another. Second, filtering and template matching technologies are employed to remove the disturbances from the background and other objects in the image. Finally, coordinate transformations are used to reconstruct the poses of the robot and the object when they are moving. The experiment results are presented to show the feasibility and the effectiveness of the algorithm.
