计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (29): 53-55.

• 学术探讨 • 上一篇    下一篇

基于分块-空间聚类的图像配准算法

龚咏喜1,刘 瑜1,谢玉波2,3,杨 平4,邬 伦1   

  1. 1.北京大学 遥感与地理信息系统研究所,北京 100871
    2.武汉大学 测绘遥感信息工程国家重点实验室,武汉 430079
    3.华北计算技术研究所,北京 100083
    4.广州市房地产测绘所,广州 510030
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-10-11 发布日期:2007-10-11
  • 通讯作者: 龚咏喜

Image registration algorithm based on blocking-spatial clustering

GONG Yong-xi1,LIU Yu1,XIE Yu-bo2,3,YANG Ping4,WU Lun1   

  1. 1.Institute of RS & GIS,Peking University,Beijing 100871,China
    2.State Key Lab of Information Engineering in Surveying Mapping and Remote Sensing,Wuhan University,Wuhan 430079,China 3.North China Institute of Computing Technology,Beijing 100083,China
    4.Real Estate Surveying and Mapping Office of Guangzhou,Guangzhou 510030,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-10-11 Published:2007-10-11
  • Contact: GONG Yong-xi

摘要: 模板匹配算法应用广泛,但不能判断配准结果是否正确,也无法比较不同像对配准结果的准确程度。提出无变形、无旋转情况下分块-空间聚类的图像配准算法,将基准图分块在参考图上配准从而获得基准图的多个配准位置,并对这些位置进行空间聚类从而计算基准图的最后配准位置,并评估配准质量。试验表明该算法配准准确度高,能够正确评估配准质量并比较不同像对配准结果的准确程度。

关键词: 图像配准, 分块匹配, 空间聚类, 配准质量评估

Abstract: While widely used,algorithm of template matching can neither judge correctness of the result nor compare accuracy of registration results of different pairs of images.Image registration algorithm based on blocking-spatial clustering is presented while images have no distortion or rotation,in which the standard image is divided into several sub-blocks to register with the reference image,positions of standard image from sub-blocks are clustered to get the final registration position and the registration quality is evaluated.Experiments show that the algorithm not only features high accuracy but also can evaluate the registration quality correctly and compare accuracy of the different registration results.

Key words: Image registration, block matching, spatial cluster, registration quality evaluation