电子地图色域映射的SOM神经网络方法

Gamut Mapping of Electronic Map Based on SOM Neural Network

颜色的设备相关性导致电子地图在跨媒介再现时容易出现色彩变形以及地理信息错误传输等问题.本文分析了电子地图色域的特征,提出了一种基于SOM(Self-Organizing Maps)神经网络的色域映射方法.针对神经网络方法没有顾及电子地图色域映射过程中的色彩空间权重以及神经网络中邻域函数的各向同性假设不适用于地图色域这两个问题,提出了改进方法.将本文方法与ICC(International Color Consortium)感知再现、绝对色度再现色域映射方法进行了对比,结果表明本文方法能够很好的保持地图整体色差,能够提高电子地图颜色复制精度.

Device-dependency of color introduces such problems as color distortion and error geographic information transmission when electronic maps represent in multiple-medias. Based on the analysis of the characteristics of map gamut,a gamut mapping method based on SOM（ Self-Organizing Maps） neural network is proposed. In this method,the weight factor of map color is considered and the neighborhood function,which is anisotropic in map color,is modified. The experiments indicate that our proposed method performs better than the ICC（ International Color Consortium） perceptual intent and absolute colorimetric intent on keeping overall color difference. This method can also improve the reproduction accuracy of electronic map colors.