Star cameras provide high-accuracy attitude determination for optical satellites, enabling precise geometric positioning of imagery without ground control points. However, affected by launch vibrations and thermal variations, the on-orbit geometric distortion differs significantly from ground measurement results, degrading attitude determination accuracy. Traditional methods typically rely on a physical model to characterize distortion, suffering from low accuracy at the detector edges. This article proposes a novel on-orbit geometric calibration method based on a block pointing model and angular distance consistency (BPMADC). Based on geometric distortion characteristics, the pointing calibration models for the main and local blocks are established. Using the consistency of stellar angular distances, the distortion parameters are solved without interference from attitude parameters. An order-staged solution and weight optimization are proposed to improve calibration accuracy. The calibration accuracy of star camera B of the Luojia3-02 satellite is improved from 0.352 pixels to 0.214 pixels. Furthermore, distortions at the detector edges are effectively calibrated. Validation data acquired on October 2, October 12, November 9, and November 26 are used to evaluate the calibration accuracy and stability of the camera parameters across different stellar regions and measurement times. The improvement rates are 48.249%, 26.856%, 44.049%, and 41.547%, respectively. Furthermore, experimental results demonstrate that the proposed method effectively improves geometric calibration accuracy and supports high-precision attitude determination for optical satellites.
Ning Zhang, Yanli Wang, Pin-Xi Liu et al.· IEEE Transactions on Geoscie...· 0 citations
On-orbit relative radiometric calibration (RRC) is a fundamental prerequisite for quantitative remote sensing analysis and high-level product generation. Although side-slither maneuvers provide a robust means for calibration, existing methodologies are often constrained by regularization inaccuracies, uneven distribution of gray-scale samples, and limited adaptability to complex multisensor architectures. This article proposes a unified RRC framework to overcome these challenges. The process begins with a prior-guided optimization method for data regularization, which operates independently of linear features or edges. To handle the inherently uneven distribution of gray-scale samples across natural scenes, an adaptive clustering-based method is implemented to estimate calibration coefficients. This approach ensures stable performance across the observed effective dynamic range, particularly in gray-scale ranges where samples are sparse. Furthermore, the framework incorporates a detail-aware strategy to achieve high-precision calibration across the full field of view (FOV). Specifically, the virtual steady reimaging (VSRI) model is first leveraged to achieve rigorous spatial alignment of identical ground features across multiple sensors. Based on this precise geometric alignment, the radiometric calibration is subsequently anchored to an optimal reference radiometric state. This approach effectively eliminates cross-chip inconsistencies while preserving structural details. Validation using side-slither and push-broom data from the Intelligent Remote Sensing Satellite-1 (IRSS-1), Luojia3-02 (LJ3-02), and Ziyuan-1F (ZY-1F) satellites demonstrates the effectiveness of our proposed approach across diverse scenes, sensor architectures, and spectral bands. Comparative analyses show that the proposed method achieves overall superior performance over four state-of-the-art methods in removing striping artifacts and maintaining radiometric fidelity.
Tao Peng, Ru Chen, Qianyu Wu et al.· IEEE Transactions on Geoscie...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
