一种微光与红外图像彩色融合方法

Color fusion method for low-level light and infrared images

微光夜视仪和红外热像仪这两种图像传感器互补的成像特性可以使它们在几乎任何条件(昼/夜/烟/雾)下观察场景。针对微光与红外图像,提出了一种彩色融合算法,可以使融合图像有相对自然真实的颜色感觉。首先用中心-环绕拮抗彩色融合算法产生彩色源图像,然后在YCbCr颜色空间中让源图像的直方图与参考图像的相匹配。为了增强融合图像的对比度,可以先用灰度融合图像代替亮度分量,然后进行直方图匹配。仿真结果表明文中提出的方法可以使融合图像接近自然真实的颜色感觉,易于分辨识别目标,从而提高观察者或机器视觉系统的工作效率,增强对总体形势的意识能力。

The complementary nature of low- le vel light image intensifiers and thermal infrared cameras makes them possible to observe the scene under almost any conditions （day/night/smoke/fog）. A color fusion method of low- level- light and infrared images was introduced to render multiband imagery in comparative realistic color appearance. Firstly, the center- surround opponent- color fusion scheme was used to yield color imagery. Next, the histograms of the image were matched to those of the reference image in YCbCr color space. This transfered the natural color characteristics of daylight imagery into multispectral night vision images. In order to enhance the contrast of the fused imagery, a grayscale fused image was used to replace the luminance component of the multiband imagery and then implement the histogram matching. The experiment results demonstrate the proposed method is effective in displaying the fused imagery in natural colors and popping out targets. This fused data can help observers or machine vision systems to process this type of images efficiently, therefore reducing detection and recognition times and improving situational awareness.