Positioning Using Star Images and a Combined Image Processing and Optimization Algorithm

The celestial navigation is one of the oldest navigation methods for determining the position, which is done by observing different celestial bodies such as stars. With the advancement of technology and the emergence of accurate imaging sensors, this method has entered a new stage, so that by processing images of stars and analyzing their data, the exact position of an observer can be determined in space or on land. This paper presents a novel method for determining the geographic location of a mobile system using an image of visible stars in the local sky at a specific time. The proposed method does not require any prior knowledge about the observer's position or orientation. Initially, a grayscale image of the sky is either generated using the Sky Map function in MATLAB, based on the SAO star catalog, or acquired directly from a star sensor. Spatial and frequency-domain features are extracted from the image and compared with features from a set of synthetic sky images corresponding to various geographic locations. A genetic algorithm is employed to optimize the matching process through a tailored fitness function. The key innovation of this research lies in the development of an algorithm capable of accurate localization using only a sky image, without reliance on auxiliary sensors or initial position data. The proposed approach is robust to noise and image rotation and is applicable across the entire Earth's surface. Simulation results confirm the effectiveness and high accuracy of the method for vision-based autonomous navigation.