Improving Geometric Performance for Imagery Captured by Non-Cartographic Optical Satellite: A Case Study of GF-1 WFV Imagery
| Autor: | Mingjun Deng, Guo Zhang, Deren Li, Qingjun Zhang, Kai Xu |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2018 |
| Předmět: |
Earth observation
010504 meteorology & atmospheric sciences Computer science Science 0211 other engineering and technologies 02 engineering and technology 01 natural sciences Data acquisition rigorous sensor model (RSM) Calibration geometric performance wide-field view (WFV) camera 021101 geological & geomatics engineering 0105 earth and related environmental sciences Block (data storage) Pixel Orientation (computer vision) non-cartographic Process (computing) Global change satellite image General Earth and Planetary Sciences Satellite rational function model (RFM) Cartography GF-1 |
| Zdroj: | Remote Sensing Volume 10 Issue 6 Remote Sensing, Vol 10, Iss 6, p 971 (2018) |
| ISSN: | 2072-4292 |
| DOI: | 10.3390/rs10060971 |
| Popis: | Numerous countries have established their own Earth observing systems (EOSs) for global change research. Data acquisition efforts are generally only concerned with the completion of the mission regardless of the potential to expand into other areas, which reduces the application effectiveness of Earth observation data. This paper explores the cartographic possibility of images being not initially intended for surveying and mapping, and a novel method is proposed to improve the geometric performance. First, the rigorous sensor model (RSM) is recovered from the rational function model (RFM), and then the system errors of the non-cartographic satellite&rsquo s imagery are compensated by using the conventional geometric calibration method based on RSM finally, a new and improved RFM is generated. The advantage of the method over traditional ones is that it divides the errors into static errors and non-static errors for each image during the improvement process. Experiments using images collected with the Gaofen-1 (GF-1) wide-field view (WFV) camera demonstrate that the orientation accuracy of the proposed method is within 1 pixel for both calibration and validation images, and the obvious high-order system errors are eliminated. Moreover, a block adjustment test shows that the vertical accuracy is improved from 21 m to 11 m with four ground control points (GCPs) after compensation, which can fulfill requirements for 1:100,000 stereo mapping in mountainous areas. Generally, the proposed method can effectively improve the geometric potential for images captured by non-cartographic satellite. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
| Nepřihlášeným uživatelům se plný text nezobrazuje | K zobrazení výsledku je třeba se přihlásit. |