研究目的
To improve the accuracy and efficiency of star centroiding methods for star sensors by developing a fast Gaussian fitting method (FGF) that approximates the solution of the Gaussian fitting algorithm (GF) in a closed-form, making it suitable for real-time applications.
研究成果
The proposed fast Gaussian fitting method (FGF) significantly improves the efficiency of the Gaussian fitting algorithm (GF) without reducing its accuracy, making it suitable for real-time applications. The algorithm performs better than existing star centroiding methods in terms of accuracy, robustness, and efficiency, as verified by both simulated star images and star sensor images.
研究不足
The fast Gaussian fitting method (FGF) requires the SNR of each pixel involved in the calculation to be close enough to 1, which may not always be satisfied in real star images due to the uncertainty of noise and star intensity. Additionally, the algorithm's performance may be affected by the distortion of the optical system in real star sensors.
1：Experimental Design and Method Selection:
The study involves developing a fast Gaussian fitting method (FGF) to approximate the solution of the Gaussian fitting algorithm (GF) in a closed-form. The methodology includes performing the FGF twice to calculate the star centroid, first to roughly calculate the Gaussian parameters and noise intensity, and then to accurately calculate the star centroid using the noise intensity from the first step.
2：Sample Selection and Data Sources:
Simulated star images and star sensor images are used to verify the performance of the algorithm. The star images are generated to simulate the images obtained by a star sensor, each of 31 × 31 pixels and includes only one star spot.
3：List of Experimental Equipment and Materials:
The experiments are carried out on the MATLAB 2016b software platform running on an Intel Core i5-7500T 2.7 GHz processor.
4：7 GHz processor.
Experimental Procedures and Operational Workflow:
4. Experimental Procedures and Operational Workflow: The proposed algorithm is tested on a set of simulated star images with different characteristics, including different centroid locations, Gaussian radii, noise levels, and brightness levels. The accuracy of the algorithm is analyzed by the star centroiding error, defined as the Euclidean distance between the calculated centroid and the true centroid.
5：Data Analysis Methods:
The performance of the proposed algorithm is compared with existing star centroiding methods, including the center of gravity algorithm (CG), the weighted center of gravity algorithm (WCG), the Gaussian fitting algorithm (GF), and the Gaussian analytic method (GA). The total time consumption of each algorithm on 10,000 star images is counted and compared.
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
