多视匹配MVLL算法及其在ADS40线阵影像中的运用

基金项目: 国家863计划资助项目(2006AA12Z154)；武汉大学测绘遥感信息工程国家重点实验室开放研究基金资助项目；辽宁工程技术大学地理空间信息技术与应用实验室基金资助项目(2007007)；矿山空间信息技术国家测绘局重点实验室开放研究基金资助项目

详细信息

MVLL Multi-Image Matching Model and Its Application in ADS40 Linear Array Images

¹Institute of Surveying and Mapping,Information Engineering University,66 Middle Longhai Road,Zhengzhou 450052,China;²Geomatics and Applications Laboratory,Liaoning Technical University,47 Zhonghua Road,Fuxin 123000,China;³Remote Sensing and Geographic Information System Researching Center,Beijing University,5 Yiheyuan Road,Beijing 100871,China

Funds: 国家863计划资助项目(2006AA12Z154)；武汉大学测绘遥感信息工程国家重点实验室开放研究基金资助项目；辽宁工程技术大学地理空间信息技术与应用实验室基金资助项目(2007007)；矿山空间信息技术国家测绘局重点实验室开放研究基金资助项目

摘要: 基于传统单立体VLL匹配算法,从多视觉立体匹配角度出发,研究了一种新的多视影像匹配MVLL算法。ADS40线阵影像多视匹配试验表明,MVLL算法在相似特征、遮挡特征等难点匹配问题上具有显著优势。
- 多视 /
- 匹配 /
- MVLL /
- 线阵 /
- 特征
Abstract: Based on the traditional single-stereo VLL matching algorithm,MVLL,which is a multi-view image matching algorithm appropriate for the matching of linear-array images,is proposed.MVLL algorithm can integrate the information of multiple overlapping images to avoid blind information area,and thus has a high successful matching rate.Algorithm analysis and experimental results show that,compared with traditional single stereo image matching algorithms,MVLL algorithm can achieve prominent performance in the matching of both similar features and occluded features.
- multiple-view /
- matching /
- MVLL /
- linear-array /
- feature