New software optimizes Webb Telescope optics to improve exoplanet imaging

A team of researchers has created a software optimization process to improve direct imaging of exoplanets using the James Webb Space Telescope, Ars Technica reported. The process, named Physics-based Learning for Imaging Planets or FLIP, optimizes a physical model of the telescope hardware to isolate faint planetary signals from background noise. [1][3][4]

The FLIP system focuses on the Near Infrared Camera instrument aboard the observatory, according to Ars Technica. The software models the optical hardware using 106 possible parameters. It applies gradient descent, a mathematical technique borrowed from artificial intelligence neural network training, to adjust these parameters on graphics processing units. [1]

The space telescope is designed to perform infrared astronomy and features high-resolution instruments capable of viewing faint or distant objects. By optimizing the physical optics model, the software helps compensate for slight mechanical instabilities in the telescope structure, Scientific American reported. [5][6]

As FTMQ Science reported earlier, the James Webb Space Telescope has previously been utilized for deep-space studies, including measuring heavy metal concentrations near the central black hole of the Milky Way and observing early dark matter structures. The new software refinement provides astronomers with improved observational precision without requiring physical modifications to the orbiting observatory. [1][6][9][10]