Abstract
A least squares reconstruction technique is examined for determining flow-field densities from optical data. Nonintrusive optical methods have long been used for flow visualization; however, the goal of this work is to devise mathematical techniques with which optical data can be used for quantitative flow measurement. The ill-posedness of density computation from interferogram measurements is recognized as a serious limitation in direct inversion methods. Here, least squares techniques employing compactness constraints are developed to avoid the difficulties encountered in traditional approaches.