The most complete catalog of space objects has appeared – stars, galaxies, and quasars. Researchers can study it through a special interface or in the form of a table.
A team of astronomers from the University of Hawaii has created the world’s largest three-dimensional astronomical catalog of stars, galaxies, and quasars. The team used data from the panoramic survey telescope and the UH or Pan-STARRS1 (PS1) rapid response system. The survey is the world’s largest multicolor optical survey, covering three-quarters of the visible sky. Astronomers have used new computational tools to decipher which of the 3 billion objects are stars, galaxies, or quasars. For galaxies, the software also produced estimates of their distances.
The resulting 3D catalog is available as a high-level scientific product. Its size is about 300 GB, and scientific users can access the catalog through the MAST CasJobs SQL interface or download the entire collection as a table.
Astronomers performed publicly available spectroscopic measurements that provide the final classification of objects and distances and passed them on to the artificial intelligence (AI) algorithm. The AI process was key in helping the team understand how to accurately define the same properties across different dimensions of objects’ colors and sizes. This AI method with a “feeding neural network” achieved an overall classification accuracy of 98.1% for galaxies, 97.8% for stars, and 96.6% for quasars. The accuracy of estimating the distance to galaxies is almost 100%.
Study lead author Robert Beck. “Using a state-of-the-art optimization algorithm, we used a set of nearly 4 million lights to train the neural network to predict source types and distances to galaxies.”