High-accuracy location algorithm of planetary centers for spacecraft autonomous optical navigation

Abstract

This study provides an on-board image processing algorithm for a spacecraft optical navigation sensor system. We propose a high-precision planetary center localization algorithm that provides effective information for subsequent spacecraft navigation measurement considering the geometric properties and special texture features of the planetary image. First, we combine a morphological opening with a minimum spanning circle method to preprocess an image. Second, a gradient direction partial area effect method in the edge region of interest is proposed for high-precision subpixel edge extraction. This study combines the law of edge energy distribution with a gradient direction partial area effect on planetary edge detection. A nonlocal filtering method is proposed to improve the results under complex conditions. Finally, the center coordinates of the planet are obtained by fitting the circular arc edges using least squares. The simulated and real images show that our algorithm can obtain high-precision subpixel edges and center coordinates for regular planetary images, with strong robustness to planetary images with noise, complex background, and rich texture.