Robust Through-the-Wall Radar Image Classification Using a Target-Model Alignment Procedure

Abstract

A through-the-wall radar image (TWRI) bears little resemblance to the equivalent optical image, making it difficult to interpret. To maximize the intelligence that may be obtained, it is desirable to automate the classification of targets in the image to support human operators. This paper presents a technique for classifying stationary targets based on the high-range resolution profile (HRRP) extracted from 3-D TWRIs. The dependence of the image on the target location is discussed using a system point spread function (PSF) approach. It is shown that the position dependence will cause a classifier to fail, unless the image to be classified is aligned to a classifier-training location. A target image alignment technique based on deconvolution of the image with the system PSF is proposed. Comparison of the aligned target images with measured images shows the alignment process introducing normalized mean squared error (NMSE) ≤ 9%. The HRRP extracted from aligned target images are classified using a naive Bayesian classifier supported by principal component analysis. The classifier is tested using a real TWRI of canonical targets behind a concrete wall and shown to obtain correct classification rates ≥ 97%.

Multipath Exploitation in Non-LOS Urban Synthetic Aperture Radar

“…To analyze the effects of (11) in the image domain, and see how it relates to the traditional point spread function (PSF) [17], [21], we make certain assumptions. Consider deriving the traditional PSF from (3) in the absence of noise and interference, i.e., , assuming zero path loss and under unit reflectivity, i.e., (by definition of the PSF).…”

“…More importantly, the PSF by definition is evaluated for a single point target return [21], whereas the CPSF operates on the returns from several assumed point like multipath targets. Theoretically the PSF is a spatial transfer function defined at the origin convolved with a spatial delta function at the target coordinates [21], [25]. The convolution is already computed in (14).…”

“…The PSF captures the range resolution via and takes into account oversampling in downrange and crossrange, if they were employed. With similar convention, the CPSF specialized for the rectangular function is (21) where is only defined for the multipath and is given by otherwise (22) We have all the necessary terms to compute the interference term in the CPSF specialized for the rectangular function, and is given by (23) Numerical results of the rectangular PSF and CPSF are presented in Section IV. In particular, it is shown that in certain cases the interference term is spatially spread in both downrange and crossrange, and its magnitude is more or less insignificant when compared to the sidelobe levels of the original PSF.…”

Through Wall Radar Classification of Human Micro-Doppler using Singular Value Decomposition Analysis

“…Previously developed signal processing techniques that have also been applied to TTW imaging include the MUSIC algorithm [12], compressed sensing [13], image alignment procedures [14] and spatial filtering [15]. The focus of these techniques is often on the maximization the signal-to-noise ratio (SNR) of the desired target or removal of the significant primary wall reflection.…”

Performance Analysis of Radar Target Recognition Using Natural Frequency: Frequency Domain Approach

“…The reflected signal of the radar target can be simulated using scattering analysis of the radar target [36][37][38][39][40][41][42][43][44][45][46][47][48][49]. In addition to radar detection and tracking, there have been mamy studies on radar target recognition [3,7,[50][51][52][53][54][55][56][57][58][59][61][62][63][64][65][66][67][68][69][70]. The expressions of recognition, classification and identification have been used interchangeably.…”