The problem of object recognition in digital images is one of the more important topics for many industrial and military applications of automatic vision systems. In addition to noise effects, the identification of imaged objects is greatly hindered by perspective modification of the object's shape and other optical-geometric effects such as the apparent position and scale (magnification). The application of a new algorithm designed to automatically assess orientation, position, and scale of objects portrayed in digital images is discussed. The algorithm, which utilizes a scene description based on image edges and the related Hough transform, also performs object-background separation. The obtained procedure is tested with excellent results using synthetic and real images of flying objects. The purpose is to exploit new techniques and algorithms for automatic target detection or tracking.