Speckle Space–Time Covariance in High-contrast Imaging

Abstract We introduce a new framework for point-spread function subtraction based on the spatiotemporal variation of speckle noise in high-contrast imaging data where sampling timescale is faster than evolution timescale. One way that space–time covariance arises pupil as atmospheric layers translate across telescope aperture and create small, time-varying perturbations phase incoming wavefront. The propagation this field to focal plane preserves some covariance. To utilize covariance, our approach uses Karhunen–Loève transform an image sequence, opposed set single reference images previous applications Image Processing (KLIP) imaging. With recent development photon-counting detectors, such microwave kinetic inductance technique now has potential improve contrast when used post-processing step. Preliminary testing simulated shows can by at least 10%–20% from original image, with significant further improvement. For certain choices parameters, algorithm may provide larger gains spatial-only KLIP.