On-Orbit High-Accuracy Geometric Calibration for Remote Sensing Camera Based on Star Sources Observation

Abstract

The on-orbit calibration technology for internal and external parameters of remote sensing camera is the key to guaranteeing the imaging quality and positioning accuracy. Conventional landmark-based methods are not suitable for satellite missions where landmarks do not work nor for the on-orbit autonomous calibration of large remote sensing constellation. Hence, we introduced an on-orbit geometric calibration method based on star sources observation. By adjusting the satellite attitude, the camera and star sensor can simultaneously detect the star sources. Taking the star sources as control points, a high-accuracy "camera-star sensor" joint geometric calibration model integrating the space-time effects and optical imaging is established, which precisely decouples the deviations caused by the relativistic effects, the distortion of internal parameters, and the bias of external parameters. Furthermore, the procedures of the method are designed for on-orbit autonomous calibration. On-orbit experiments were carried out in several satellites, such as Jilin-1 ShiPin07 (JL-1 SP07). The results showed that the accuracy of internal calibration could reach 0.052" and the error of external installation matrix was within ±1.4". Therefore, the method is attractive for actualizing on-orbit autonomous geometric calibration accurately and complements the conventional methods where the availability of landmarks is hard.

High-Accuracy Real-Time Attitude Determination and Imagery Positioning System for Satellite-Based Remote Sensing

“…Xing et al toke the stars as the control points for conducting the external calibration with the Jilin-1 SP07. The error of the external parameters was within ±1.4″ [15]. These calibration-based methods can provide reliable and accurate external installation.…”

Accurate Pointing Determination of Space Targets under the Stars Background

“…Usually, taking the first order of two kinds of distortion can well fit the lens distortion. Therefore, the distortion model is expressed as follows (Chen et al, 2022):…”