Image jitter detection and compensation using a high-frequency angular displacement method for Yaogan-26 remote sensing satellite

Satellite platform jitter is an important factor restricting the imaging quality of high-resolution (HR) optical satellite images. To address the critical issue of compensation for attitude jitter in HR images, this paper proposes a steady-state reimaging model using high-frequency angular displacement data to detect and compensate for the attitude jitter of HR images. The bidirectional Kalman filter and overall weighted smoothing method helps realizing information fusion of star sensor and angular displacement sensor and obtaining the high-frequency attitude for image jitter detection. Then, the steady reimaging model is used to correct the distorted image with geolocation consistency based on a rigorous geometric model. The Yaogan-26 remote sensing satellite's distorted panchromatic images of airports, targets and calibration fields affected by platform jitter were used to validate the effectiveness and accuracy of the proposed method. The compensation results show that the proposed method can effectively improve the relative geometric quality of images affected by platform jitter, with the images' jitter distortion being clearly eliminated. Compared to the conventional compensation method that bundle adjustment with GCPs, the absolute geometric accuracy can also be improved.