Passive Three-Dimensional (3-D) Imaging and Ranging

Creation of 3-D data from 2-D imagery has vast applications, particularly in defense related areas. Two such applications of great interest to the Air Force are real-time and lightweight imaging device for seekers on board of autonomous munitions and for systems that supports navigation and tactical reconnaissance from unmanned-aerial-vehicles (UAVs) and small UAVs also known as Small Air Vehicles (SAV). These systems carry imaging payloads operating in either visible or infrared wavelengths. 3-D images provide unit commanders with a much intuitively useful representation of the area of operations than simple 2-D photos, and greatly aids automated detection and tracking of objects in the air or otherwise. UtopiaCompression Corporation (UC) and the team member Defense Research Associates Inc., (DRA) proposes (1) Novel structure-from-motion (SFM) framework (augmenting and refining existing algorithms), using a robust feature matching technique, instead of the conventional Harris corner detector, to derives structure and motion of dynamic scene containing multiple moving objects observed from a stationary/non-stationary camera; (2) The derived structure and motion parameters are used in conjunction with the metric self-calibration algorithm to obtain 3-D model, which is physically more appealing for human perception; and (3) for far-off objects particle filter is used to derive depth information.