Master thesis and internship[BR]- Master's thesis : Improvement of the optical setup of a compressive sensing imager[BR]- Internship

Abstract

Compressed sensing is an emerging field that allows for the recovery of a sparse signal from fewer measurements than permitted by the Nyquist theorem. This new paradigm can be used to create new imager architectures that are simpler, more compact, and cheaper than traditional imagers, acquiring images in a compressed manner and thereby reducing the amount of data to handle. These characteristics are appealing for potential implementation in Earth observation satellites, where size and weight are critical factors, and where the amount of collected data is substantial, with limited storage capacity and transfer rates to the ground. This master’s thesis focuses on the implementation of the optical part of a compressive sensing imager in the laboratory, with the objective of performing a particular compressive sensing reconstruction method, known as inpainting. A comprehensive review of the state of the art in compressive sensing and various imager architectures is first provided. Afterwards, the design of the instrument and the selection of all its components are extensively detailed. A digital micromirror device is used for producing incomplete, damaged images of a scene, and a camera detector records the resulting image. Subsequently, a calibration procedure for the damaged images was established to prepare them for reconstruction through inpainting. This calibration includes dark and fl at frame corrections, as well as post-reconstruction image perspective correction. The instrument’s point spread function is also measured, and a dithering method is implemented to improve its resolution. Furthermore, the pattern mask used for the reconstruction is studied and calibrated using morphological erosion.