基于静态小波分解和能量函数优化的图像拼接

Image Mosaic Based on Stationary Wavelet Decomposition and Energy Function Optimization

提出了一种基于静态小波分解和能量函数优化的全景图拼接算法.该算法首先使用静态小波分解初始图像,并利用序贯检验法对高频图像快速搜索计算粗略匹配序列,在此基础上对初始图像细致搜索计算精确匹配点;对于传统图像融合处理中采用的线性加权函数通常引起的重叠区域模糊问题,采用包含图像梯度的能量函数计算高频图像的全局最优融合因子进行融合,而低频图像则采用线性因子融合.实验表明,该算法在降低匹配运算时间的同时又能获得满意的视觉效果.

A novel method for image mosaics based on stationary wavelet decomposition and energy function optimization was presented.First,stationary wavelet decomposition and series similarity detection were used for searching the corresponding points in high frequency images.Then delicate search were made in origin image for exact matching based on these corresponding points.Fusion factors for high frequency detail images were calculated from energy function that contains image gradient, while linear weight function were used for low frequency smooth image.This conquers the blur and mackle problems in the overlap area using traditional linear weight function.Experiment results show that satisfying visual effect can be achieved using this method while matching time is also reduced in image mosaics.