摘要
为了对同一光学传感器的多聚焦图像进行融合增强以获得较高清晰度的图像,提出了一种基于中值金字塔的图像融合算法.首先,利用中值滤波和分析、合成算子得到图像的各级金字塔分解,以及各级细节子图像.然后,对顶层低频子图像系数取平均,对各级细节子图像,以局部邻域平均梯度为判据,进行图像融合.最后,利用中值金字塔反变换重构图像.实验结果表明,该方法对多聚焦图像的融合十分有效.
To obtain a higher definition enhanced-image with multi-focus images from same sensor,an image fusion algorithm based on median pyramid is proposed.First,obtain the pyramid decomposition as well as the details subimages on each level by using median filtering and analysis and synthesis operators.Then,make an average towards the coefficients of the top subimage.The fused image of subimages on each level can be gained according to the local average gradient criterion.Finally,the reconstructed image can be obtained though inverse transform of median pyramid.The experimental results are evaluated and indicate that the algorithm is effective to process the multi-focal images.
出处
《微电子学与计算机》
CSCD
北大核心
2008年第9期165-167,共3页
Microelectronics & Computer
基金
江苏省图像处理与图像处理重点实验室开放课题(ZK206008)