低照度夜视成像的自然感彩色化及增强方法  

Natural-appearance Colorization and Enhancement for the Low-light-level Night Vision Imaging

作　　者：朱进 李力 金伟其 李硕 王霞 拜晓峰 ZHU Jin1 , LI Li1 , JIN Wei-qi1, 2, LI Shuo1 , WANG Xia1 , BAI Xiao-feng2(1MOE Key Laboratory of Optoelectronic Imaging Technology and System, School of Optics and Photonics, Beijing Institute of Technology, Beijing 100081, China; 2 Science and Technology on Low-Light- Level Night Vision Laboratory, Xi ＇an 710059, Chin)

机构地区：[1]北京理工大学光电学院光电成像技术与系统教育部重点实验室,北京100081 [2]微光夜视技术重点实验室,西安710059

出　　处：《光子学报》2018年第4期159-168,共10页Acta Photonica Sinica

基　　金：国家自然科学基金(No.61405013);陆装十三五基金(No.6140414050327);微光夜视技术重点实验室基金(No.BJ2017001)资助~~

摘　　要：低照度成像系统输出一般都是黑白视频,为了获得更好的夜视实际应用效果,提出了一种基于YUV空间色彩传递的低照度视频图像彩色化及增强的亮度拉伸色彩传递方法.该方法借鉴了双波段色彩传递自然感彩色融合方法,由灰度图像及其负片图像构成初始彩色图像,并对亮度通道进行自适应亮度拉伸,在UV通道进行参考图像的色彩传递,实现灰度图像的自然感彩色化和增强.通过与其他基于色彩传递的彩色化方法比较,亮度拉伸色彩传递方法对参考图像和源图像的相似程度要求较低;选取几幅适当的典型场景彩色参考图像,可对绝大多数场景取得较好的彩色化效果,具有很好的场景环境普适性.同时可以看出,该方法高效,对比度高,颜色协调性好,色彩自然,更有利于人眼的观察感知,对于低照度夜视成像效果提升效果明显.该方法已在硬件平台上实时应用,可在无需增加硬件资源的基础上,有效地应用于低照度夜视成像.The output video of the low-light-level solid-state imaging devices are always gray.For better low-light-level imaging applications,a natural-appearance colorization and enhancement method named Luminance Stretching Color Transfer（LSCT）for grayscale video images using color transfer is proposed.A two-channel natural-appearance color fusion method is refered to in the LSCT method.In order to achieve the natural-appearance colorization and enhancement,firstly,the pre-colorized image is obtained by combining the grayscale image with its negative image.Following this,an adaptive luminance stretching is performed and color of the reference image is transferred in the YUV color space.As compared with other methods based on color transfer,the LSCT method is less affected by the degree of similarity between the reference image and the original grayscale image.It means that relatively good results may be achieved for most scenes with an appropriate reference image.Thus,the LSCT method has better environmental adaptability.The comparisons reveal that the LSCT method is high efficient and its colorized results appear more natural in respect to human perception with better contrast and color harmony.Moreover,the LSCT method has been implemented in real time on hardware platforms.Therefore,it can effectively improve the effect of human observation to apply our method in the lowlight-level imaging without increasing any hardware costs.

关 键 词：低照度 彩色夜视 色彩传递 图像增强 实时成像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]