基于相似场景的低照度监控图像增强  

LOW-ILLUMINATION SURVEILLANCE IMAGE ENHANCEMENT BASED ON SIMILAR SCENES

作　　者：朱婧雅 王中元[1] 

机构地区：[1]武汉大学计算机学院国家多媒体软件工程技术研究中心,湖北武汉430000

出　　处：《计算机应用与软件》2015年第1期203-205,210,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(60970160;61070080)

摘　　要：为了使低照度监控图像增强后的效果近似于自然光照下的图像,更有利于辨识,对基于相似场景的低照度监控图像增强方法进行研究。首先获取良好光照下,与低照度监控图像场景相似的图像,并建立图像库,从中选择最合适的图像作为参考图像;然后通过直方图匹配的方法,以参考图像为标准,对低照度图像进行增强,并用迭代增强的方法不断提高图像的清晰度;最终得到增强后的低照度监控图像。实验结果表明,该方法改善了图像的视觉效果,提高了图像的亮度和对比度。In order to make the low-illumination surveillance images similar to the natural lighting images after enhancement and more con- ducive to identifying, we study the similar scenes-based enhancement method for low-illumination surveillance images. First, we collect the images in good lighting which are similar to the scenes of low-illumination surveillance images, and use them to create an image library, and choose the most appropriate one as the reference image. Then we use histogram matching method to enhance the low-illumination image with the reference image as standard. Furthermore, we employ iterative enhancement method to constantly improve the clarity of the image, and finally get the enhanced low-illumination surveillance image. Experimental results show that the method perfects the visual effect of the images and improves their brightness and contrast as well.

关 键 词：低照度图像 彩色图像增强 直方图匹配 迭代增强 视频监控 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]