Low-Rank Matrix Recovery Approach for Clutter Rejection in Real-Time IR-UWB Radar-Based Moving Target Detection

Abstract

The detection of a moving target using an IR-UWB Radar involves the core task of separating the waves reflected by the static background and by the moving target. This paper investigates the capacity of the low-rank and sparse matrix decomposition approach to separate the background and the foreground in the trend of UWB Radar-based moving target detection. Robust PCA models are criticized for being batched-data-oriented, which makes them inconvenient in realistic environments where frames need to be processed as they are recorded in real time. In this paper, a novel method based on overlapping-windows processing is proposed to cope with online processing. The method consists of processing a small batch of frames which will be continually updated without changing its size as new frames are captured. We prove that RPCA (via its Inexact Augmented Lagrange Multiplier (IALM) model) can successfully separate the two subspaces, which enhances the accuracy of target detection. The overlapping-windows processing method converges on the optimal solution with its batch counterpart (i.e., processing batched data with RPCA), and both methods prove the robustness and efficiency of the RPCA over the classic PCA and the commonly used exponential averaging method.

On the Applications of Robust PCA in Image and Video Processing

“…This application is the most investigated one. Indeed, numerous authors used RPCA problem formulations in applications such as background/foreground separation [4], [208], [223], background initialization [255], [258], moving target detection [241], motion saliency detection [47], [300], [332], motion estimation [238], visual object tracking [168] [276], action recognition [126], key frame extraction [60], video object segmentation [130], [153], [197], [317], [319], video coding [45], [46], [110], [331], video restoration and denoising [142], [334], [109], [318], [176], video inpainting [142], hyperspectral video processing [96], [42], and video stabilization [68].…”

“…Thus, several authors designed RPCA formulation for videos taken by a fixed color CCD camera (in most of the cases), but also by hyperspectral camera [257], by camera trap [97], [98], [246] and by aerial camera [79], [80], [81], [82]. Furthermore, dedicated methods also exist for infrared cameras [241] and RGB-D cameras [267], [132]. For the environments which present dynamic backgrounds, illumination changes, camera jitter, etc., many modified RPCA approaches have been designed according to the following very popular background modeling challenges:…”

“…A weighting matrix W or a transformation τ can also be used in the constraint of the minimization. In literature, numerous publications used the robust PCA formulation by improving its computational efficiency and its adequacy to the concerned application in (1) signal processing applications like in satellite communication [150], seismology [67], [53], speech enhancement [125] [290], Synthetic-Aperture Radar (SAR) imaging [127], [307], [240], [159], [155], [329], [23], [24], [95] and direction-of-arrival tracking [167], [61], (2) computer vision applications like in image processing, video processing and 3D computer vision as developed in Bouwmans et al [28], (3) computer science applications such as the detection traffic anomalies [191] [216], and (4) astronomy for auroral substorm detection [304] and exoplanet detection [101], [217]. A full list of publications of RPCA in these different applications is available at the DLAM website 1 .…”

Extended Target Echo Detection Based on KLD and Wigner Matrices

“…The principal difference between the two approaches lies in that coherent integration takes into account the phase information of the target, so that the accumulation effect is better than the noncoherent integration. A typical coherent integration algorithm is the moving target detection (MTD) algorithm [2], which is implemented by fast Fourier transform (FFT) in a single range unit. Although the most effective way to achieve target detection at low SNR is to extend the target accumulation time, long-term observations can cause the target echo envelope to present across range unit (ARU) and Doppler frequency migration (DFM) phenomena [3,4], which seriously restrict the long-term accumulation performance of the target.…”

Total Variation Regularization Term-Based Low-Rank and Sparse Matrix Representation Model for Infrared Moving Target Tracking

“…Fortunately, low-rank and sparse representation is an effective tool that is employed in many applications of computer vision, such as fore-and background separation [26][27][28][29][30][31][32], fabric defect inspection [33], face recognition [34], act recognition [35], and so on. Its basic principle is that the foreground occupies a small number of pixels and that the background images are linearly related in consecutive frames, meaning that the fore-and background patches can be treated as a low-rank matrix and a sparse matrix, respectively.…”