双波段红外图像的对比度增强融合

Contrast-enhanced fusion of dual-band infrared image

为了抑制在处理双波段红外图像融合时低对比度图像细节信息丢失导致漏警的问题,在基于WLS边缘保持平滑的基础上,提出了一种多尺度对比度增强的融合算法。通过边缘保持平滑,提取不同尺度的细节分量,从而实现对比度在不同空间频率上的增强;为了进一步消除图像光晕现象,采用非下采样方向滤波,使多分辨率变换同时具有方向性。实验结果表明,本算法能够使融合图像包含输入图像更多的互补信息,图像细节更加清晰,较大地提高了红外弱小目标的对比度。

In order to suppress the false dismissal phenomenon caused by the detail information loss of the image with low contrast when dealing with dual- band infrared image fusion,a fusion algorithm on the basis of edge- preserving smoothing via WLS was proposed. Through edge- preserving smoothing,detail components in different scales were extracted to enhance the contrasts in different spatial frequencies. In order to further eliminate the halo phenomenon,non- subsample directional filter was adopted to make multi- resolution transform has directional character. Experimental results show that proposed algorithm can make the fusion image contain more complementary information of input images and clearer image details,and the contrast of infrared dim target was enhanced largely in meanwhile.