研究目的
To improve the geolocation accuracy of the GF-3 SAR satellite by compensating for atmospheric path delay using a sophisticated tropospheric model based on real meteorological data.
研究成果
The integral tropospheric model achieves millimeter-level accuracy and improves GF-3's geolocation accuracy by several decimeters compared to the static model. This enhancement is significant in areas with complex weather conditions. Future work should incorporate precise corner reflectors and address other error sources like solid Earth tides to achieve sub-meter accuracy.
研究不足
The use of natural targets as GCPs instead of corner reflectors introduces elevation errors. Satellite trajectory errors from real-time orbit products (nominal accuracy of 10 m) and control point extraction errors (within one pixel) affect accuracy. The study is limited to specific test areas and time periods, and the integral model requires extensive meteorological data, which may not be universally available.
1：Experimental Design and Method Selection:
The study compares three tropospheric models (simplified static, Saastamoinen (SAAS), and integral models) for estimating zenith tropospheric delay (ZTD) and applies them to GF-3 SAR geolocation. The integral model, based on real meteorological data, is used to compensate for atmospheric path delay.
2：Sample Selection and Data Sources:
Meteorological data from ECMWF ERA-Interim reanalysis (2016 data, 6-hour intervals, 37 vertical layers) and zenith path delay (ZPD) products from IGS stations (BJFS and URUM) are used. GF-3 SAR images from various modes (SL, FS2, SS, QPS2) acquired in 2017 are analyzed, with ground control points (GCPs) manually extracted from Google Earth.
3：List of Experimental Equipment and Materials:
GF-3 SAR satellite data, ECMWF meteorological data, IGS ZPD products, and computational tools for data interpolation and model calculation.
4：Experimental Procedures and Operational Workflow:
For model accuracy analysis, ZTD is estimated using the three models and compared to IGS ZPD data via bilinear interpolation. For geolocation assessment, atmospheric path delay is compensated in GF-3 images using different models, and range accuracy is evaluated against GCPs.
5：Data Analysis Methods:
Root-mean-square error (RMSE) and difference values (DIFF) are calculated to compare model accuracies. Geolocation errors in range are analyzed statistically.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
