基于方向梯度局部方差的分层遥感图像融合

为了充分利用各种遥感图像的信息．本文对以往图像融合技术进行了简要的概述，分析了各种方法的优缺点．给出了一种新的基于方向梯度局部方差的金字塔图像融合方法。将各源图像进行金字塔分解，得到多分辨率描述．建立方向梯度金字塔．采用局部方差准则融合源图像的各分解层．再进行金字塔分解的逆变换，得到融合图像。此外．利用信息熵等准则和其他融合方法进行比较．实验结果表明用此方法能得到更好的融合效果。

Hierarchical Fusion of Remote Sensing Image Based on Local Deviation of Direction Gradient

In order to adequately use all kinds of remote sensing image information, some important image fusion techniques, analysis features and disadvantages of these methods are summarized, a novel hierarchical fusion technique for remote sensing image is presented based on local deviation of the direction gradient. Original images are decomposed into multi resolution representations by using pyramid transform. Gradient pyramids are obtained by gradient operators. Original images are fused in each level of different input resources based on local deviation. The resultant image can be arrived by performing inverse wavelet transform. These methods by computing their entropy and square error are compared. Experimental results show that more satisfactory effect is obtained.