Stereo Vision-Based Motion Estimation for Space Point Targets

Abstract

This paper investigates the problem of accuracy improvement for the motion estimation of space point targets in the early detection stage from high-orbit infrared remote sensing satellites. Firstly, according to the imaging characteristics of the high-orbit satellite, the paper establishes the imaging mapping models and obtains the simulated images as the inputs. Then, the three-dimensional position recovery equations of the target are constituted through the relationship of imaging mapping models between two satellites. Next, the target trajectory is fitted by cubic spline interpolation, and a differential algorithm solves the target velocities. Finally, considering line-of-sight (LOS) measurement error in actual observation, the Savitzky-Golay filter is used to smooth the estimated target positions and velocities. Simulation results show that the proposed method has an obvious advantage over the extended Kalman filter (EKF) method in real-time tracking accuracy at an early target detection stage under the same design parameters and errors.