面向影像地图色域映射算法研究

A Gamut Mapping Algorithm for the Photomaps

为了使影像地图跨媒介传递色彩失真最小,本文提出一种面向影像地图色域映射算法。首先,使用双边滤波对影像地图进行预处理,分解出低频和高频图像,根据人眼空间域视觉感知理论,对低频图像采用保护彩度裁切法,对高频图像采用保护明度压缩法,最后融合两幅映射后图像并再次映射至目标色域,使用图像结构相似度和S-CIELAB色差模型进行客观评价,并与HPMINΔE法、SGCK法及Bala提出算法进行比较。结果表明,面向影像地图色域映射算法在图像结构相似度和图像色差上都比其他算法稍好。因此,面向影像地图色域映射算法能提高影像地图的映射质量,可同时保护影像地图影像细节和矢量符号色彩。

To make the photomaps reproduction color distortion minimum during cross- media transmission,a gamut mapping algorithm was proposed for the photomaps. Bilateral filtering was used to preprocess the photomaps,which was decomposed into the low frequency and the high frequency image,according to the theory of the human visual perception in the spatial domain,Chroma protection clipping algorithm was used for the low frequency image,lightness protection compression algorithm was used for the high frequency image,then the processed images were merged,finally the merged image was mapped to the target gamut once again. Structure similarity and the S- CIELAB color difference model was used to evaluate objectively,the data was compared with the minimum color difference method,SGCK and the algorithm proposed by Bala. The results indicated gamut mapping algorithm for photomaps was slightly better than other algorithms on image structure similarity and color difference. Gamut mapping algorithm for photomaps could improve the quality of the photomaps mapping,at the same time could protect the photomaps image details and color of vector symbols.