摘要
针对现有图像拼接方法的缺点提出一种全自动图像拼接算法,该算法综合考虑了图像拼接的精度和速度。利用迭代法自动求得图像的最佳分割阈值,二值化原始图像,经过标记、筛选等处理后快速提取出边缘重叠图像匹配过程中所需的特征模板。通过提取出的特征模板利用基于灰度的方法确定图像的最佳拼接线,从而实现图像正确、平滑的无缝拼接。实验结果证明了该算法的有效性。
By analyzing the available disadvantages of the image mosaic, a fast automatic image mosaic algorithm is proposed. This algorithm takes into account the precision of image Mosaic as well as the speed of it. A best value which is used to segment the image is get automatically by the iterative algorithm. Then the image is changed to a kind of image which has only two values （0 or 255 ）. After the processing of labeling and filtration , rapidly work out the feature blocks of the overlapping images that will be necessary in the matching later. In order to finish matching the images correctly and smoothly, the best place of the line is got based on the way of gray. The experiment shows that the proposed algorithm has good capabilities.
出处
《现代防御技术》
北大核心
2007年第6期110-114,共5页
Modern Defence Technology
关键词
图像拼接
特征区域
模板匹配
图像融合
image mosaic
feature blocks
template matching
image inosculation
