Fast restoration of smeared navigation images for asteroid approach phase. (November 2020)

• Record Type: Journal Article
• Authors: Lu, Xiaoxuan; Zhu, Shengying; Liang, Zixuan
• Is Part Of: Acta astronautica. Volume 176(2020)
• Page Start: 287
• Page End: 297
• Publication Date: 2020-11
• Subjects: Optical navigation -- Feature extraction -- Image restoration -- Asteroid exploration; Astronautics -- Periodicals; Outer space -- Exploration -- Periodicals; Astronautics; Periodicals; 629.405
• DOI: 10.1016/j.actaastro.2020.06.024
• Languages: English
• ISSNs: 0094-5765

Abstract

Abstract: Optical navigation is one of the key technologies to enhance the autonomy of spacecrafts. However, optical images may undergo severe motion degradation during asteroid approach phase, which adversely affects the accuracy of navigation. To guarantee the navigation accuracy for future asteroid missions, this study proposes a novel restoration method for smeared navigation images captured by onboard optical sensors. By taking advantage of the intensity distribution of reference stars, a fast estimation and refinement method for motion degradation is drawn from star stripes directly without the aids of other sensors. On this basis, a complete processing pipeline is put forward to recover the navigation images in real space environment. The proposed method does not take the assumption of constant angular velocities and enables the capability of dealing with complex motion degradation that may occur over the long exposure time for capturing faint space objects. Simulations are conducted using both synthetic and real images. The proposed method can have real-time performance while maintaining good restoration accuracy, and thus may have the potential for onboard implementation. Highlights: A novel restoration method for smeared navigation images is proposed. A fast method to estimate motion degradation is derived only using star stripe information. The proposed method does not reply on the assumption of constant angular velocities. Fault detection via the chi-squared test is designed to reject fault measurements which may occur occasionally. The processing pipeline performs well for both unresolved and resolved targets.