Star trackers are the most accurate attitude-determination systems for satellites. Its operation is based on the identification of the stars present in the field of view of a small telescope and the determination of the astronomical coordinates of the centre of the image. To allow star identification, the star tracker uses a catalogue of stars whose size and properties primarily depend on the expected accuracy and the field of view. Identifying the stars present in the field is a difficult and time-consuming process for which there exist many different algorithms. It is necessary to benchmark many of these algorithms in terms of accuracy, computational burden, and size of the star catalogue, among other figures of merit. A specially demanding phase of attitude determination is the case in which there is no prior information on the orientation of the satellite, as is the case when it is just released from the rocket. This is referred to as Lost-in-Space (LIS) and requires the identification of the star field from scratch. This Final Degree Thesis is devoted to the exploration of some of the leading algorithms usable with limited computational power expected to be available in small satellites. The codes, which must be valid for LIS and nominal operations, will be written in MatLab and their performance measured through the standard MatLab tools for such purposes. Our project is particularly relevant in the context of the development of attitude determination and control of PhotSat, a 12U CubeSat with astronomical applications currently being developed in the Institute of Space Studies of Catalonia.