An optical satellite Self-Geolocation method based on Multi-Angle observations

• Observing an area from different observe angles to realize self-geolocation. • The method is completely independent of calibration fields and easy to implement. • The accuracy calculation formula was derived and validated using simulation data. • Real data experiments proved the accuracy is close to the traditional method. • The execution cost of the algorithm is much lower than traditional algorithms. Traditional optical satellite geometric positioning methods have certain limitations, including the high costs of calibration field construction, the time required, and significant susceptibility to weather conditions at calibration sites. In the present era of ever-increasing satellite constellation deployments, such conventional methods struggle to meet operational demands. Consequently, self-geolocation technologies are receiving more attention. This paper presents a self-geolocation method that completely eliminates the dependency on ground control points. The method uses images acquired during different orbits as inputs and leverages the geographic positioning consistency of corresponding points to calculate a sensor installation matrix without any need for control points. Initial experiments using simulated data revealed the method to be feasible. Subsequent validation using real data from the Luojia 3–02 satellite showed that the positioning accuracy was comparable to traditional method, but the calculation cost was significantly lower. The method will find applications in satellite constellations and long-term precision maintenance missions.