图像融合是一种重要的增强图像信息的方法.提出了一种基于多尺度分解的像素图像融合方法.利用小波变换按照不同融合规则及融合算子构造融合图像对应的小波系数.通过对可见光图像与红外图像进行融合的实验比较表明,以像素的局部能量为准则的融合效果更好,既可避免传统融合规则的信息损失,又提高了融合图像的空间分辨率和清晰度,融合图像更符合人的视觉特性,也有利于机器视觉.

Image fusion is one of the important techniques to enhance image information. A novel pixel-level image fusion scheme is presented based on multiscale decomposition. The wavelet transform is used to construct the wavelet coefficients of fused image using multiple operators according to different fusion rules. Comparison of fusion results of visible image and infrared image is given. It shows that the proposed method based on local area energy of each pixel can achieve better fusion performance than other usual fusion rules, avoiding losing information and improving the resolution and the special quality of image. Then the fused image is more suitable for human vision or machine perception.