Aerospace and Electronic Systems Magazine July 2017 - 48

Automatic Target Recognition in Missing Data Case

Figure 14.

Six models, including sedans (a, b, c), SUVs (d, e), and a van (f) [3].

However, an unknown class is not included in the ATR performance evaluation. The vehicles are divided into two groups, the
training data set (116 sedans, 91 SUVs, and 61 vans) and the testing data set (115 sedans, 91 SUVs, 61 vans).
Table 1 compares the ATR performance of images formed via
BP with those formed using the hybrid approach for full, 30%, and
50% missing data cases.
Again, the hybrid imaging approach offers superior performance in classification error rate in all data loss cases examined.

REFERENCES
[1]
[2]

[3]

CONCLUSIONS
The hybrid high-resolution imaging method for wide-angle SAR
has been shown to provide enhanced ATR performance of civilian
vehicles in full data cases, as well as missing data cases compared
with other data-independent cases such as the BP method. The
hybrid imaging approach yields high-quality images that in turn
provide simple and representative target features. In addition, the
hybrid pose estimation and simple vertical CSV profiles facilitate
SAR image registration as a preprocessing step for target classification. The combination of the hybrid high-resolution SAR imaging,
the efficient pose estimation, the simple CSV profiles, and the reliable target classification method makes it possible to distinguish the
three civilian vehicle types (sedan, SUV, and van), even when up to
50% of the data is corrupted. The hybrid imaging method, consisting of SIAA, FIAA, and SLIM, yields unbiased estimates as well as
sidelobe suppression, which results in enhanced ATR performance
compared with the BP method as shown in Table 1.

[4]

[5]

[6]

[7]

[8]

IEEE A&E SYSTEMS MAGAZINE

Dudgeon, D. E., and Lacoss, R. T. An overview of automatic target
recognition. The Lincoln Laboratory Journal, Vol. 6, 1 (1993), 3-10.
Karlsson, J., Rowe, W., Xu, L., Glentis, G. O., and Li, J. Fast missingdata IAA with application to notched spectrum SAR. IEEE Transactions on Aerospace and Electronic Systems, Vol. 50, 2 (Apr. 2014),
959-971.
Dungan, K. E., Ash, J. N., Nehrbass, J. W., Parker, J. T., Gorham, L.
A., and Scarborough, S. M. Wide angle SAR data for target discrimination research. In Proceedings of SPIE Conference on Algorithms for
Synthetic Aperture Radar Imagery 8394, May 2012.
Carrara, W. G., Majewski, R. M., and Goodman, R. S. Spotlight Synthetic Aperture Radar: Signal Processing Algorithms. Norwood, MA:
Artech House, 1995.
Yardibi, T., Li, J., Stoica, P., Xue, M., and Baggeroer, A. B. Source
localization and sensing: A nonparametric iterative adaptive approach
based on weighted least squares. IEEE Transactions on Aerospace and
Electronic Systems, Vol. 46, 1 (Jan. 2010), 425-443.
Glentis, G. O., and Jakobsson, A. Efficient implementation of iterative
adaptive approach spectral estimation techniques. IEEE Transactions
on Signal Processing, Vol. 59, 9 (Sept. 2011), 4154-4167.
Tan, X., Roberts, W., Li, J., and Stoica, P. Sparse learning via iterative
minimization with application to MIMO radar imaging. IEEE Transactions on Signal Processing, Vol. 59, 3 (Mar. 2011), 1088-1101.
Gorham, L. A., and Moore, L. J. SAR image formation toolbox for
MATLAB. In Proceedings of SPIE Conference on Algorithms for Synthetic Aperture Radar Imagery 7699, Apr. 2010.

JULY 2017

Table of Contents for the Digital Edition of Aerospace and Electronic Systems Magazine July 2017