一种新的多传感器图像融合方法

提出了一种基于小波变换的改进图像融合方法和一种图像融合评价指标。对小波分解后的高频分量使用图像区域内象素的最大绝对值作为该区域中心象素的活性测度以得到融合图像的高频分量，对分解后的低频分量通过度量图像区域质量来选择该区域中心象素从而确定融合图像的低频分量，最后进行小波重构得到融合图像；文中还提出了用交互方差作为图像融合客观评价指标的方法。实验结果表明，该融合方法得到的图像清晰度和对比度都得到了较大的提高，是一种有效的图像融合算法；并且用交互方差作为图像融合的客观评价指标是可行的。

A Novel Multi-Sensor Image Fusion Method

An improved image fusion algorithm based on wavelet transform and a novel evaluation criterion for image fusion are presented. The image is decomposed into high-frequency image and low-frequency image firstly. The maximum absolute value within the window of high-frequency image is used as an activity measure associated with the center pixcl to get the fused high-frequency image. The quality of the blocks within the low-frequency image is measured to determine the center pixel of the fused low-frequency image. The inverse transformation is performed to get the fused image at last. The proposed algorithm is verified to be effective for image fusion because the information, the clarity and the contrast of the fused image are all improved. At the same time, the mutual variance is proposed as one of the objective evaluation criteria for the fused image and is proved to be feasible in image fusion.