Study and implementation of a High-Dynamic Range (HDR) imaging algorithm for a space telescope

Abstract

The sun's corona exhibits a very high brightness gradient. The use of HDR (High dynamic range) imaging technique is an asset at improving the quality of the resulting image. By taking successive images with different exposure times, it is possible to enhance the contrast of the image. Short exposure time images being adapted for very bright parts of the corona but inadequate for dark parts. Whereas long exposure time images are adapted for dark parts but inadequate for bright parts where saturation may occur. Combining images of different exposure time allows to take advantage of both and is at the basis of HDR techniques. The combination of images requires a perfect alignment of the sun's coronal structures. Images of different exposure time are taken at successive instants which causes misalignment due to the relative motion of the objects in the scene. An image alignment algorithm based on contours of intensity in the image has been developed. It allows the alignment of multi-exposure images with a pixel precision. The composition of multi-exposure images allows to get a higher dynamic range to capture the details of the corona of the sun. Special care must be taken to avoid problems due to saturation and noise of the different images as well as unwanted objects in the scene. An algorithm for the composition of an HDR image has been developed. It is applied in the case of the PROBA-3 satellite project: a new type of external coronagraph. As this is a future project, data had to be simulated for this study. Rather than giving one method to achieve the goal, it is a study of the different possibilities and will serve as a guide of possible techniques for further implementation. It is part of the internship made at the Centre spatial de Liège (CSL).