Star identification in night sky images using mobile phone camera

Stars in the night sky are dim light sources. This makes it difficult to capture the details of the night sky using small, low aperture camera sensors, such as those used in mobile phones. Therefore, processing must be done to correctly identify stars in an image captured from a mobile phone, such that the resultant image depicts clearly visible stars with a high signal-to-noise ratio. Existing literature primarily focus on astrophotography using Digital Single-Lens Reflex camera devices (DSLR), which have larger sensors and are kept stationary using mounts such as a tripod. There are very few research works pertaining to astrophotography using mobile phone cameras, due to the lower processing power and the inconvenience to keep a mobile phone stationary during image acquisition. Techniques involving machine learning need a large dataset of preprocessed photos of the night sky. Such a dataset is not publicly available for research. In this work, we develop a novel approach for star identification, enhancement and image stacking using image processing techniques, in order to derive plausible results for night sky images taken from a smartphone sensor. Image enhancement pipelines for three different scenarios have been designed, namely for pure night sky images, night sky images with peripheral objects and burst shot images of the night sky. The results observed are promising and the proposed approach has a potential to be employed for night sky enhancement in mobile phones.