Aircraft identification by unions of ISAR images

Abstract

We offer an algorithm that can identify aircraft categories from Inverse Synthetic Aperture Radar (ISAR) images that use both the radar reflection pulse shape, which includes the duration or size of the radar pulse that is reflected, and the Doppler shifts of different parts of the aircraft caused by rotational motions of the aircraft as it maneuvers. We investigated the practicality of determining which of seven different aircraft categories a radar return indicates. The object of this research is to very quickly tell from an ISAR return how an aircraft compares to the seven different categories where the aircraft is in any position of a prescribed holding pattern. We propose a new method in which we compare each ISAR image to unions of images of the different aircraft categories. This method gave us results that are superior to the results we obtained in [8].

Algorithms for radar image identification and classification

“…The results obtained with the both proposed classification methods are compared testing the same aircraft model sets. The second proposed union algorithm [10] shows itself to be much more successful in correct and efficient classification than the algorithm described in [8] and [9]. The percent of correctly classified images with the union method is three to thirty times higher than with the method proposed in [8] and [9].…”

“…Category 4 had approximately 2.2% error, Category 1 had less than 1% error and Category 5 had less then 1/3 % error. The union algorithm [10] is far superior not only in accuracy but also in speed to the results obtained with classification method described in [8] and [9].…”

“…In the unions of ISAR images classification algorithm [10] we compare the unknown model with the union images of every aircraft category present in the database that are shown in Figure 10. We tested each individual image from 360 images set from categories 1 to 5, 180 from category 6 and 60 from category 7 according to the proposed union classification algorithm.…”

“…The proposed algorithm for union of ISAR images classification is very simple and it is described in [10]. It makes a single comparison of the captured ISAR image to the union of images of all positions in a holding pattern of each aircraft category.…”

“…Section 3 describes the two proposed classification algorithms. Section 4 provides details of our experimental results and compares our proposed methods described in [8], [9], [10] and Section 5 concludes our discussion.…”