Researchers have created a new method for detecting distant exoplanets by computationally removing the overwhelming glare of their host stars from telescope images. The technique, which relies on a detailed virtual replica of the James Webb Space Telescope’s infrared camera, has demonstrated the ability to reveal planets that were previously obscured in existing observations and achieved two to 15 times greater detection sensitivity compared to conventional approaches in tests with known planets.
The breakthrough, called Physics-based Learning for Imaging Planets (FLIP), works by creating a precise physical model of how the James Webb Space Telescope’s Near Infrared Camera operates. Rather than attempting to improve the telescope’s hardware, the method uses advanced computational techniques to identify and account for tiny imperfections and subtle variations that affect observations. By simulating millions of possible factors that could influence the telescope’s performance, researchers can model the appearance of a star’s light with high precision and then subtract that modeled glare from actual images.
The development of FLIP involved collaboration between researchers at Caltech, Northwestern University, and NASA’s Jet Propulsion Laboratory. Lead author Rodrigo Ferrer-Chavez, a graduate student at Northwestern, worked under the mentorship of astronomer Jason Wang and Caltech’s Katie Bouman, who specializes in computational imaging. The team’s approach represents a shift from earlier attempts to use machine learning to predict starlight patterns, instead leveraging optimization techniques to adjust millions of variables and refine their simulation of the telescope.
According to the researchers involved, the iterative process of determining what the telescope actually does during observations produces increasingly cleaned images, eventually revealing the faint signals from distant planets. Ferrer-Chavez noted that this represents the first demonstration that minor imperfections and instabilities in the telescope need not be limiting factors in directly imaging exoplanets. A paper describing the method was published online November 4.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI