一种新的多传感器彩色图像融合方法

A novel multi-sensor color image enhanced fusion algorithm

针对多传感器彩色图像的融合问题,提出了一种基于高斯同态滤波和YIQ变换的融合方法。对源图像进行YIQ变换,并设计一种动态高斯同态滤波器,在频域内处理各图像的亮度分量,在保持图像彩色信息的同时增强图像细节。运用不同的小波变换规则,对亮度分量和色度分量分别进行融合。将各分量的融合结果逆变换至RGB空间,得到融合图像。实验分析表明:新算法得到的融合图像无论在视觉效果还是在客观评价方面,均优于小波变换方法和Laplace塔型方法,具有令人满意的融合效果。

Focusing on the problem of muhi-sensor color image fusion, a novel fusion algorithm based on Gauss homomorphic filtering is proposed. The initial images are transformed to the YIQ color space, and a dynamic Gauss homomorphie filter is introduced to process their luminance components in the frequency range, so that the color information is reserved while the image detail is enhanced. Different rules of wavelet transform are employed to fuse the luminance component and ehrominanee components respectively. The fused components are inverse ly transformed to RGB color space to obtain fused image. Experimental analysis demonstrate that, the proposed algorithm brings better vision effect and subjective evaluation to the fused image than Wavelet Transform and Laplace Tower and its fusion effect is satisfied.