Parallel manipulator tracking with occlusion based on improved mean-shift

Abstract

An improved mean-shift algorithm is proposed and applied to track a moving parallel manipulator with partial occluded. At first, an improved mean-shift algorithm based on color-texture features is introduced. Then a factor reflecting the severity of occlusion is defined to judge whether the end-effector of parallel manipulator is occluded. At last, the end-effector is divided into several sub-blocks and each sub-block is tracked independently. The sub-block with the largest confidence decides the location of the end-effector in next frame. The experimental results demonstrate that it is effective and efficient.