GOAL:

The goal of this work is to provide an empirical basis for research on star detection and tracking .

To this end, we have collected hand-taken star frames from 3 different smartphones models and videos from a timelaps camrea (taken over the night)

Star-tracking algorithm are essential for orientation computation of devices such as autonomous platforms such as nano-satellites, drones, vessels and automobiles.

Using commercial-off-the-shelf (COTS) mobile devices such as smartphones for star trackign is especially challenging because it requires robustness and should work continuously and autonomously. This dataset can be used for exploring and testing of such algorithms.

DataSet:

The frames in this data set were all taken by hand and on different conditions ( clear and cloudy nights) , therefore different glare is expected .

In S9 folder : for each frame we added a text file with the stars coordinates x,y, radius (in pixels). This text file was generated for each image using simple image processing algorithm that search for the brightest pixels (as connected components) in each frame and compute the center of each star in sub-pixel.

Folders:

S9:

This folder contains frames from Galaxy s9 smartphone.

The camera resolution and FOV parameters can be found here:

http://www.samsung.com/global/galaxy/galaxy-s9/specs/

S6:

This folder contains frames from Galaxy s8 smartphone.

The camera resolution and FOV parameters can be found here:

https://www.gsmarena.com/samsung_galaxy_s6-6849.php

S7:

This folder contains frames from Galaxy s7 smartphone.

The camera resolution and FOV parameters can be found here:

https://www.gsmarena.com/samsung_galaxy_s7-7821.php

starVideo_ timeLapsCamrea:

This folder contains the time laps camera star videos. those videos were taken through the night 3 minutes appart .

this is the camera we used in this experiment: https://home-fix.com/products/brinno-tlc200-pro-hdr-time-lapse-video-camera

DOWNLOADS: