Abstract  In this paper the shape information contents of a morphological vector descriptor, called pecstrum (pattern spectrum), are investigated. The pecstrum is then used for aircraft recognition and classification. The pecstrum is a simple vector descriptor which provides information on the way the area of the object is distributed from the fine details to its bulky contents. Although some of its properties have already been reported [3], [4], [14], [23], the use of the pecstrum as a classification tool has not been given appropriate emphasis. At the beginning of the paper some introductory material on mathematical morphology and the pecstrum is presented for the reader who is not familiar with the relevant terminology. Next the shape information which the pecstrum conveys is analyzed and its classification properties are considered. New concepts such as the pecstral space and the cumulative pecstral transformation are introduced and explained. The performance of the pecstrum in certain recognition problems is also examined. The concept of B-shapiness is redefined and the relation between the pecstrum and the ratio area/perimeter² is established. The pseudopecstrum is then introduced and its information contents and classification properties are compared with those of the conventional pecstrum. The use of pecstrum in estimating object orientation is also addressed. Finally, the recognition and classification capabilities of the pecstrum are tested using a large number of binary objects (airplanes). The performance limit of the pecstrum for efficient object classification, as the size of the objects decreases, is examined and the factors which affect this limit are discussed. The classification results are compared with those obtained using invariant moments.