Searcharxiv⌕ Search

arXiv subjects

Sandrine J. Thomas

Publications and source records attributed to Sandrine J. Thomas.

20 records · Page 2Linked to original sources

Experimental study of a low-order wavefront sensor for the high-contrast coronagraphic imager EXCEDE

The mission EXCEDE (EXoplanetary Circumstellar Environments and Disk Explorer), selected by NASA for technology development, is designed to study the formation, evolution and architectures of exoplanetary systems and characterize circumstellar environments into stellar habitable zones. It is composed of a 0.7 m telescope equipped with a Phase-Induced Amplitude Apodization Coronagraph (PIAA-C) and a 2000-element MEMS deformable mirror, capable of raw contrasts of 1e-6 at 1.2 lambda/D and 1e-7 above 2 lambda/D. One of the key challenges to achieve those contrasts is to remove low-order aberrations, using a Low-Order WaveFront Sensor (LOWFS). An experiment simulating the starlight suppression system is currently developed at NASA Ames Research Center, and includes a LOWFS controlling tip/tilt modes in real time at 500 Hz. The LOWFS allowed us to reduce the tip/tilt disturbances to 1e-3 lambda/D rms, enhancing the previous contrast by a decade, to 8e-7 between 1.2 and 2 lambda/D. A Linear Quadratic Gaussian (LQG) controller is currently implemented to improve even more that result by reducing residual vibrations. This testbed shows that a good knowledge of the low-order disturbances is a key asset for high contrast imaging, whether for real-time control or for post processing.

astro-ph.IM↗

Computer vision applications for coronagraphic optical alignment and image processing

Modern coronagraphic systems require very precise alignment between optical components and can benefit greatly from automated image processing. We discuss three techniques commonly employed in the fields of computer vision and image analysis as applied to the Gemini Planet Imager, a new facility instrument for the Gemini South Observatory. We describe how feature extraction and clustering methods can be used to aid in automated system alignment tasks, and also present a search algorithm for finding regular features in science images used for calibration and data processing. Along with discussions of each technique, we present our specific implementation and show results of each one in operation.

astro-ph.IM↗