摘要
针对图像配准,提出了一种高效的配准方法,结合小波变换与传统的配准方法,检测形状的轮廓,利用小波分解后的系数分层提取图像形状信息,并进行由粗到细的逐次配准,然后利用最小距离函数对两轮廓进行判别,减少了计算量,得到了高精确的配准结果,有效地提高了配准效率。
Coming up with image registration, the authors put forward a efficient way of registration which reduced computation, acquired precise registration, and improved efficiency of registration. Combining traditional registration with wavelet trans-formation, the authors successively registered image registration from coarseness to thinness, using coefficient stratification decomposed by wavelet. Finally, the authors distinguished 2 profiles through function of minimum distance.
出处
《科技创新与生产力》
2010年第5期86-87,共2页
Sci-tech Innovation and Productivity
关键词
图像配准
小波变换
形状匹配
image registration
wavelet transformation
shape matching
