融合云模型和比重阈值的唐卡破损区域分割

Damaged region segmentation of Thangka using combining cloud model and proportion threshold

针对唐卡破损区域分割问题,提出了融合云模型和比重阈值提取破损区域的算法。首先获取RGB彩色图像的灰度图、HSV空间的V分量、YIQ空间的Y分量、LUV空间的L分量,合并这四个分量得到融合图像;其次利用云模型过滤融合图像,得到云过滤图像;然后计算云过滤图像的比重度并进行局部比重阈值分割,得到分割结果 1;再分块粗分割融合图像得到分割结果 2和分割结果 3;最后合并分割结果 1、2和3得到分割结果 4,对分割结果 4去除过分割区域,得到最终分割结果。实验结果表明,与一维最大熵法、OTSU算法、数字形态学算法等算法相比,该算法的分割效果较好并具有一定的稳定性。

According to the damaged regions segmentation of Thangka problem, the paper proposes an algorithm with which the damaged regions can be segmented by combining the cloud model and proportion threshold. Firstly, grayscale image of RGB color image, V component of HSV space, Y component of YIQ space, and L component of LUV space can be obtained,to combine the four component will generate the fusion image. Secondly, when filtered fusion image by cloud model, the cloud filtering image be acquired. Thirdly, to calculate the proportion image with clouding filtering image and then to segment proportion image by local proportion threshold, we will obtain segmentation result 1. Fourthly, segmentation result 2, segmentation 3can be obtained by block segmenting algorithm on fusion image. Finally, to merge segmentation result 1,2 and 3, then we will gain segmentation result 4, when removed over segmentation of segmentation result 4, final segmentation result can be received.The experimental results show that the algorithm is more effective to segment damaged region than one-dimensional maximum entropy method, the OTSU algorithm, digital morphology algorithm etc. and has a certain stability.