关键词: 阈值分割, 图像处理, 动态环境, 色标模型, 颜色空间, 自适应

Abstract: Thresholding is a classic issue in the field of image processing and analysis, and also it is one of the most difficulties in the field. The quality and effect of thresholding benefited from the threshold dividing method, in this paper we study the recognition of color pattern model under dynamic surrounding, and current static image processing methods cannot work well for the changing vision and illumination when the robot moving, under this circumstance we put forward a adaptive threshold segmentation method, we first study utilizing different color space to recognize the target under dynamic environment, and choose the best color space, then through updating the next center point and recognition range which we estimated by the current image to reach the adaptive threshold segmentation, thus not only maximize the identified target, but also effectively shortening the threshold range estimated, we combine HSV and YCbCr color space to correct and reduce the estimated center point to improve the accuracy of our algorithm as well, at last we experiment and compare to get the good conclusion.

Key words: Keywords: threshold segmentation, image processing, dynamic environment, color pattern model, color space, adaptive