结合区域特性和非子采样SPT的图像融合方法

针对基于方向可控金字塔变换的图像融合方法中存在的缺点与不足,提出了一种基于区域特性和非子采样方向可控金字塔变换(NSSPT)的图像融合方法.首先采用NSSPT对源图像进行多尺度、多方向分解,得到低频子带、高频子带以及各方向带通子带系数;然后,针对高频和带通子带系数的选择,结合各子带图像的区域特性,给出了一种基于区域均值能量匹配的"加权平均"与选择相结合的系数融合策略;而对于低频子带系数则给出一种基于灰度均值偏差的选择与加权平均的系数选择方案,得到了融合图像的NSSPT系数.最后,经过NSSPT逆变换得到融合图像.对多组不同的源图像进行融合实验仿真,实验结果表明该方法可以避免"人为"效应或高频噪声的引入,能够获得视觉效果更佳、细节更为丰富的融合图像,其融合效果要优于基于传统的金字塔变换、小波变换以及方向可控金字塔变换的图像融合方法.

Regional Feature Self-Adaptive Image Fusion Method Based on Nonsubsampled Steerable Pyramid Transform

To conquer the weakness of existing traditional image fusion method based on the steerable pyramid transform,a novel adaptive fusion algorithm of multi-sensor images based on nonsubsampled steerable pyramid transform(NSSPT) is proposed.Firstly,the NSSPT is performed on the source images with different scales and directions,thus both the low and high frequency subband coefficients together with varieties of directional bandpass subband coefficients are obtained.Secondly,for the low frequency subband coefficients,a selection principle based on the local area difference of the coefficient’s mean value is presented,while for the high frequency subband coefficients and varieties of directional bandpass subband coefficients,a scheme based on the local area average energy combined with the weighted average scheme is presented,which is also consistent with the regional feature of the high and bandpass sub-images.Finally,the fused image is obtained by performing the inverse NSSPT on the combined coefficients.The experimental results show that the proposed approach not only can avoid the introduction of the artifacts and high frequency noise,but also can significantly outperform the traditional image fusion methods based on the pyramid transform,wavelet transform or steerable pyramid transform in terms of both visual quality and objective evaluation criteria.