Significance   At present, the mobile tracking station, with its advantages of high mobility, high flexibility, wide observation range and low networking cost, is gradually becoming an important system component in the space target monitoring network and is widely used for common-view observation and precision tracking of spatial objectives. The rapid angular velocity motion of the space debris causes the dynamic axis of the telescope to become unstable during tracking, resulting in a dynamic bias in the telescope's pointing. In particular, the operating conditions of the mobile station are complex and frequently changing, leading to deviations in the pointing accuracy provided by the telescope's encoder. However, precise pointing accuracy is a prerequisite for astronomical positioning and target identification by optoelectronic telescopes, so dynamic pointing errors in mobile stations must be corrected to ensure the accuracy of space target positioning. To address the dynamic pointing errors caused by the above factors, a pointing error correction method based on star pattern matching deviation calibration is provided.

  Methods  Firstly, the acquired images are processed by extracting the centroid coordinates of the measured stars, filtering the catalogue of calibration stars, and converting the stellar position coordinates. The Source-Extractor software is used to perform threshold segmentation, contour extraction and centroid coordinate extraction of the measured stars on the acquired images (Fig.2). The catalogue of calibration stars is filtered by the rough pointing provided by the telescope's encoder to determine the population of calibrated stars at the current pointing and the field of view. The SOFA package is called to perform a coordinate transformation of the coordinates of the calibration stars, transforming the flat position of the filtered calibration star at J2000 to the apparent position of the station, i.e. the azimuth and elevation of the calibration stars (Fig.4); Secondly, a rapid star pattern matching algorithm for star deviation calibration is used to identify the coordinates of calibration stars that match the measured stars, and take them as the theoretical positions. The star pattern matching algorithm for the first frame is based on classical triangle matching to improve robustness and adds star deviation calibration features to speed up the construction of feature triangles (Fig.5), the matching of subsequent frames uses the plate constants derived from the first frame to calculate the celestial coordinates of the measured star and compares them with the calibration stars in the filtered catalogue to determine whether the difference is within the tolerance limits (Fig.6); Finally, the pixel coordinates of all the measured stars are brought into a mathematical model of star deviation calibration to fit and calibrate the telescope pointing (Fig.1).

  Results and Discussions   In order to verify the effectiveness of the pointing correction method and the accuracy of the correction, a verification experiment of the pointing correction algorithm was carried out on a 400 mm aperture photoelectric telescope at a station located at 125.4443° longitude and 43.7907° latitude. The relevant telescope parameters are given (Tab.1). The experimental results demonstrate that the correction period of a single frame is about 2.2 s when a set of sequential images is acquired to correct the optical centre pointing, and the amount of correction generally stabilises from the 10th frame onwards. The pointing corrections were applied to a group of sub-sky regions with a typical distribution of the total sky area, and the mean pointing error is increased from 124.24″ to 4.97″ (Fig.9) and the standard deviation is increased from 41.50″ to 4.76″ before the correction (Fig.10). The telescope was pointed at the standard source Polaris, and the corrected photocentre pointing is 1.776″, which is different from the theoretical value, that is, the correction accuracy of this method is better than 1.8″ (Fig.11).

  Conclusions   The above experiments show that the pointing correction method based on star pattern matching deviation calibration is effective in improving the pointing accuracy of the station's telescope. The method is reliable and accurate in complex mobile station conditions and is suitable for correcting the pointing of the mobile station's telescope. In addition, the correction process is independent of the telescope's frame configuration, so it can be applied in pointing corrections for telescopes with different frame configurations, such as equatorial or geostrophic.