Fast and Cheap Color Image Segmentation for Interactive Robots

Vision systems employing region segmentation by color are crucial in real-time interactive mobile robot applications such as object tracking (e.g. the ball in RoboCup so ccer) and various forms of robot/human interaction. Traditionally, systems employing real-time color-based segmentation are either implemented in hardware, or as very specific software systems that take advantage of domain knowledge to attain the necessary efficiency. However, we have found that with careful attention to algorithm efficiency fast color image segmentation can be accomplished using commodity image capture and CPU hardware. our paper describes a system capable of tracking several hundred regions of up to 32 colors at 30 Hertz on general purpose commodity hardware. The software system is composed of three main parts; a novel threshold classifier, a region merger to calculate connected components, and a separation and sorting system to gather various region features and sort them by size. A key to the efficiency of our approach is a new method for accomplishing color space thresholding that enables a pixel to be classified into one or more of up to 32 colors using only two logical AND operations. A niave approach could require up to 192 comparisons for the same classification. The algorithms and representations are described, as well as descriptions of three applications in which it has been used.