基于双树复小波的图像镶嵌算法

Image mosaic algorithm based on dual-tree complex wavelet

为保证图像的无缝镶嵌以及清晰度与平滑度一致,提出一种图像镶嵌算法.基于双树复小波近似平移不变性和良好的方向分析能力,在确定初始重叠区域后,利用双树复小波分解结构的层次对应匹配法和Lev-enberg-Marquardt优化算法,逐层求解.求解时先对小波系数进行阈值化处理,在保证匹配精度的同时提高运算速度.结果表明,该算法实现了全景图的自动生成,全景图在视觉上看不出拼接痕迹,达到无缝拼接的效果.

A image mosaic algorithm was developed to keep seamless mosaic of images and image distinction to be in harmony with smooth degree, which considered the nearly perfect properties of shift invariance and good directional selectivity of the dual-tree complex wavelet. Using dual-tree complex wavelet level correspondence match method and Levenberg-Marquardt optimization algorithm, the wavelet coefficients were processed by threshold from approximate to detailed level. Computational speed was improved as well as match precision guaranteed. The results show that the panorama production, which the trail of seam can not be observed visually, is implemented by the algorithm. The panorama keeps seamless.