Bookshelf

This thesis develops a camera-supported aiding system for Satellite Laser Ranging (SLR). The aiding system calculates corrections to the orbit that is used to track a space object with laser ranging. While a priori orbits for official laser ranging targets already have a high precision, this is not the case for space debris objects. For debris objects, a priori orbits are generated by radar measurements that have high inaccuracies. There are existing image processing systems that detect objects in night sky images. However, these systems often have a high runtime and are not suited to be used in real-time for laser ranging. Many of the used systems operate with a fixed camera position, which leads to a low visibility of the moving objects. This work proposes a robust aiding system that leverages existing, publicly available a priori coordinates to move the camera along the object's path. By concentrating the illumination on a small group of pixels, the system increases its sensitivity to small objects, thus enhancing detection capabilities. To verify this approach, different satellites and debris objects are tracked with a camera in this work. This work shows that objects can be detected from radar predictions as old as three weeks, with offsets of up to 3°. By enhancing the aiding system with a priori coordinates, debris objects with a radar cross-section of up to 0.1 squared meters are detected. This work optimizes the object detection and astronomic reduction algorithms, such that their total runtime is below 0.1 seconds on an image of 2000 x 1300 pixels. For the popular astronomic reduction tool astrometry.net , this work demonstrates the scope of possible optimizations. This work integrates the tool as a library and replaces slow components of its solve-field executable with a custom, C++ implementation. This way, a runtime of 20 ms for astronomic reduction is achieved, improving upon the stated runtime in similar image processing systems.