Analysis of Camera's Images Influenced by Varying Light Illumination for Design of Color Segmentation *

摘要

Cameras are excellent sensors in robotic systems. However, a camera being extreme sensitive to illumination variation restricts its capability. In this paper, color image seg-mentation of a camera influenced by varying light illumination is studied for extending camera’s capability in the future. The influence of an image captured by camera due to varying light illumination is analyzed. The components of image data represented by colors make the influence of varying light illumination visible for conveniently analyz-ing. In the analysis, the change rules of the YUV components due to varying illumina-tion is induced to formulate a database consisted of the set of boundaries. Based on the database, an adaptive algorithm for color segmentation against unstable lighting is pro-posed. The development of the algorithm is summarized by a design procedure. Practical experiments demonstrate the performance of the proposed algorithm. In addition, the adaptive algorithm against uncertain light illumination is included.