Development and assessment of a complete ATR algorithm based on ISAR Euler imagery

Abstract

The Euler decomposition, when applied to the polarization scattering matrix, attempts to extract phenomenological information about the scattering target. Because the Euler parameters constitute a more physically relevant set of parameters than the traditional HH-VV ISAR representations, they have potential to improve ATR performance. The Euler parameter's usefulness in target recognition, however, is effected by several layers of signature variability. Unfortunately, many of the variability layers are often omitted in a typical ATR study. A complete ATR algorithm was therefore developed that allows for all layers of variability and requires no previous knowledge of the target's position, orientation, or average reflectivity. The complete ATR algorithm was then used to assess the effectiveness of Euler ISAR imagery in target recognition when all layers of variability are considered. The general approach and sub-methods used to construct the complete ATR system will be presented, including the methods to determine the targets orientation, registration, and to compare it to a library of pre-rendered target images. Finally, the performance of the Euler parameters in target recognition using the complete ATR algorithm will be presented.

Imaging and structural feature decomposition of a complex target using multi-aspect polarimetric scattering

“…It has been noted in Table 1 that the Huynen parameters cannot separate three scattering mechanisms for the same values of ν and γ [18,19].…”

A comparison of machine learning methods for target recognition using ISAR imagery

“…Previous work leveraged the highly accurate compact radar range data for ATR evaluation. [1,2] To that end, machine learning methods were investigated to determine which would be most effective. shows an example of ISAR images for comparison.…”