基于纹理特征库的微光图像色彩纹理传递  

Texture Library Based Color & Texture Transferring for LLL Images

作　　者：许真[1] 孙韶媛[1] 代中华[1] 郭子乾[1] 谷小婧[1] 

机构地区：[1]东华大学信息科学与技术学院,上海201620

出　　处：《红外技术》2011年第1期49-55,共7页Infrared Technology

基　　金：国家自然科学基金项目(编号:61072090;60705006)

摘　　要：对微光夜视图像进行了色彩和纹理的传递。在多次实验的基础上,从拍摄的大量微光图像里,提取不同模式的纹理块,构建了天空、道路、草地、树木四类景物的微光图像库,并结合灰度共生矩阵和Gabor滤波器提取图像纹理特征,建立其对应的纹理特征库。然后通过比对目标图像像素点纹理特征与特征库中各向量的相似性,来判定该像素所属类别,实现了微光图像的分割;最后,通过套印算法进行快速的色彩与纹理传递。实验结果表明,该算法实现了对微光图像较为准确的分割,得到了同时具有自然色彩和纹理的彩色夜视图像,更有利于人眼对目标的识别。In this paper an algorithm to transfer colors and textures to low light level night vision images is proposed. Firstly, based on lots of experiments, we extract different patterns of textures to construct a LLL image library including sky, road, grass and tree. Then we extract the texture features of each image in the LLL image library based on combination of GLCM and Gabor filter and construct the corresponding texture features library. Secondly, we classify each pixel of the original target LLL image through comparing the similarity of the pixel texture features with the texture features in the texture features library. Finally, we transfer colors and textures to the LLL image using image overprint method. The experimental results show that the color LLL images with natural colors and textures and express more harmonious visual perception.

关 键 词：图像处理 微光图像 纹理特征库 图像套印 

分 类 号：TN911.73[电子电信—通信与信息系统]