Automatic Instantaneous Attitude Estimation of Satellite Based on Optical and ISAR Image Fusion

The instantaneous attitude estimation of satellite becomes a focal topic in the increasingly significant space situation awareness area. Both the optical, inverse synthetic aperture radar (ISAR) and their fusion way have been widely applied. Due to the domain gap between two different types of data, it is difficult to make use of them for automatic attitude estimation. In view of this, this paper proposes an automatic instantaneous attitude estimation method based on optical and ISAR image fusion. The method deduced mathematical expressions of target's projected features and state parameters. Take inspiration from human joint position extraction network, a key points extraction network is proposed for automatic points extraction on both the optical and ISAR image. These extracted points are utilized to determine the projected features of target's typical structures which then be used for attitude estimation. By this means, the attitude of satellite can be estimated through optimization solved by the particle swarm optimization (PSO) algorithm. Some experiments are present to validate the feasibility and superiority of the proposed method.