Toet的夜视图像近自然色染色算法对各颜色分量使用了基于线性变换的颜色传递技术, 染色结果易出现低对比度或过饱和现象, 从而造成细节信息的损失。针对此问题, 提出了结合多尺度融合与非线性颜色传递的微光与红外图像染色算法。选用亮度与色彩分离的 YCbCr颜色空间进行染色, 其中 Y分量应用基于冗余小波变换的可增强感兴趣目标的融合方法构造得到, 对 Cb、Cr分量使用基于直方图匹配的非线性颜色传递方法实现染色。实验结果表明, 本文方法可有效防止染色过程中出现的低对比度或过饱和现象, 能获得纹理细节更清晰, 目标更显著的染色图像。

A linear color transfer technique for each component is used in the method of near natural coloration for night vision image proposed by Toet, but the phenomenon of low-contrast or supersaturation easily appears in the results which cause the loss of detail information. Aiming at the problem, this paper proposed a novel approach for coloring the low light level and infrared image using multi-scale fusion and nonlinear color transfer technique. The YCbCr color space in which chrominance and luminance are separated is adopted for coloration process. The Y component is constructed by the fusion method which is based on redundant wavelet transform and can enhance the interesting targets; the Cb and Cr component are constructed by the nonlinear color transfer technique based on histogram matching. The experimental results show that the new method effectively prevents the phenomenon of low-contrast or supersaturation in the process of coloration, and obtain the images with clear texture and prominent targets.