Research on thermopaint color image segmentation and its application in temperature recognition

A color image segmentation algorithm for thermopaint color image segmentation base on labeling region growing is proposed by analyzing the color temperature property of thermopaint and quantizing image according to average image color moment, and this algorithm is to label every region in HSI color space that has vision consistency with human eyes when region segmenting. Atter image segmenting and feature extracting, according to the color temperature relations and each corresponding disperse point constructed a continuous and smoothing three dimensions space curve, and then begin to temperature recognize through this curve. At the same time, comparing with simulating human eyes recognizing algorithm, using the recognition algorithm according to the color temperature relations table and the color temperature characteristic curve. and contrasting the results of this two methods by experiments. Experimented results show that thermopaint image recognition using image segmentation and recognition algorithm can obtain accurate detection temperature results, which can never be matched by the manual detection.