Autonomous Spacecraft Pose Estimation Method in Earth-Moon Space Based on Single Sensor

Given the constraints of geographical and temporal conditions in Earth-Moon space, the real-time groundbased navigation for spacecraft is limited and suffers from significant delays and signal attenuation issues. Therefore, optical autonomous navigation for spacecraft through the observation of Earth-Moon stars has been widely studied and verified, using visible light and infrared sensors for pose estimation. However, due to the payload weight and volume, this method does not meet the lightweight requirements of micronano satellites. This paper proposes a novel optical autonomous navigation method that utilizes a single visible light sensor to achieve spacecraft pose estimation in a distant retrograde orbit (DRO). The method identifies star points using a voting triangle algorithm to solve for attitude information, extracts the EarthMoon contours and performs elliptical fitting, and extracts the centroid using the ellipse model direct method to calculate the camera position. To verify the feasibility and accuracy of this method, single-frame pose estimation experiments were conducted using simulated images. The experimental results show that the proposed optical autonomous navigation method achieves an attitude accuracy of 9′, a three-axis position accuracy of 50 km, and an angular resolution of 0.013°.