Highly accurate image navigation and registration (INR) of data collected using remote sensors is key to satisfying mission requirements for satellite observation systems. Locating image data in a specifically defined scene has become a challenge of providing sub-pixel level knowledge of the location of the raw collected samples when computed in the designated scene. The precision INR solution presented in this study is one in which the instrument is used to collect measurements as a means of estimating the instrument line of sight attitude and thermal line of sight perturbations. The approach is applied to the GOES-R Advanced Baseline Imager (ABI) mission which collects star measurements (visible and IR) and using a Kalman filter with an order of magnitude fewer states than baseline GOES, estimates line of sight attitude and thermal drift. The algorithm computes corrections to the line of sight to compensate for thermal changes as the sample data are navigated and co-registered to the specified scene. Instrument line of sight and thermal deformation effects modeling are presented and simulation results show the attitude determination accuracy for the instrument line of sight as well as the location accuracy of samples navigated to the mission-defined scene.