测绘学报 ›› 2017, Vol. 46 ›› Issue (2): 179-187.doi: 10.11947/j.AGCS.2017.20160138

• 摄影测量学与遥感 • 上一篇    下一篇

考虑定向参数精度信息的TerraSAR-X和SPOT-5 HRS影像RFM联合定位

程春泉, 张继贤, 黄国满, 张力, 杨景辉   

  1. 中国测绘科学研究院, 北京 100830
  • 收稿日期:2016-03-28 修回日期:2016-11-10 出版日期:2017-02-20 发布日期:2017-03-07
  • 作者简介:程春泉(1972-),男,博士,副研究员,研究方向为摄影测量与遥感影像处理。E-mail:cspring@casm.ac.cn
  • 基金资助:
    国土资源部公益行业科研专项(201411119-2)

Combined Positioning of TerraSAR-X and SPOT-5 HRS Images with RFM Considering Accuracy Information of Orientation Parameters

CHENG Chunquan, ZHANG Jixian, HUANG Guoman, ZHANG Li, YANG Jinghui   

  1. Chinese Academy of Surveying and Mapping, Beijing 100830, China
  • Received:2016-03-28 Revised:2016-11-10 Online:2017-02-20 Published:2017-03-07
  • Supported by:
    Special Funds of Ministry of Land and Resources for Public Wolfere Projects (No.201411119-2)

摘要: 卫星影像的RFM模型具有传感器无关的优点,适用于多源影像的几何定位处理,但在无地面控制点条件下联合定位时存在自主定位优势影像难以发挥主导作用且求解易发散的不足。本文通过将影像的先验自主定位精度和成像线性漂移转化为像方定向参数的精度和权信息,建立考虑影像定向参数精度信息的RFM模型。以12景TerraSAR-X和6对12景覆盖面积约为18万km2的SPOT-5 HRS立体长条带影像为数据源,对两类影像定向参数先验精度配置偏差、SAR影像升降轨道方向、SAR影像数目、SAR影像分布等因素对定位精度的影响进行了系列定位试验,少量SAR与大范围HRS联合的影像自主定位平面/高程精度可达6.0 m/4.2 m。本文RFM平差模型无地面控制点定位精度和定向参数求解稳健性相对于传统模型有显著提升,是卫星影像无控制点1:10万/1:5万全球测图的一种潜在方法。

关键词: RFM, 无控制点定位, 异源影像, TerraSAR-X, SPOT-5 HRS, 全球测图

Abstract: RFM is suitable for geometric positioning with multi-source images because of the independence of sensors. However, without GCPs the geometric advantages from the geometrically prevailing images cannot be enforced in the traditional RFM-based orientation. It also leads to the apparent flaw of divergence of solving. By translating the priori information of autonomous positioning accuracy and imaging drift errors to bias compensation parameters in image space, it is rebuild a new RFM-based adjustment model which considers the accuracy and weight of the orientational parameters. A test study of positioning without GCPs which includes 12 scenes of TerraSAR-X images and 6 pairs of long-strip stereo SPOT-5 HRS images covering area of 186 000 km2 is conducted. Furthermore, the different impact factors on positioning accuracy, i.e., the accuracy of orientation parameters, the number of SAR image, orbit direction of SAR satellite and distribution of SAR images, are investigated. The results show that it can achieve positioning accuracy of 6.0 m in plan and 4.2 m in height for the case of geometrical combination of TerraSAR-X and HRS images. The proposed model is superior to the traditional RFM-based positioning with respect to accuracy and resolving stability. It is a potential approach for worldwide mapping at scale of 1:100 000 and 1:50 000 without GCPs by combination of different satellite images.

Key words: RFM, image positioning without GCP, positioning with different source images, TerraSAR-X, SPOT-5 HRS, worldwide mapping

中图分类号: