摘要
首先检测两幅图像中的角点,然后自适应地提取基准特征模板,再利用改进的基于特征强度响应空间的Hausdorff距离对基准模板进行初始匹配,最后通过区域相关法进行优化。算法不要求特征间的一一对应,也无需距离变换,实验证明这是一种快速有效的图像配准算法。
First, comers are detected in both images by SUSAN algorithm. Then, basic feature template is extracted by an adaptive strategy based on local comers density. In comer strength response space, template matching between two images is conducted based on improved Hausdorff distance. Last, registration parameters are optimized by using area correlation coefficient. Experiments demonstrate that this algorithm is fast and efficient.
出处
《计算机应用与软件》
CSCD
北大核心
2007年第8期17-18,49,共3页
Computer Applications and Software
基金
国家自然科学基金资助项目(60573179)。
