首页|Robust principal component analysis combined with top-hat transform for clutter suppression in GPR images
Robust principal component analysis combined with top-hat transform for clutter suppression in GPR images
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
Taylor & Francis
ABSTRACT As a contactless and non-invasive tool, ground penetrating radar (GPR) plays an important role in buried object detection. However, the performance of GPR images is deteriorated severely due to the clutter in radar echo signals, including echoes reflected from ground-air surface and other undesired signals. Low-rank sparse decomposition (LRSD) has been proved to be an effective tool to separate clutter and targets as low-rank and sparse components respectively. However, the properties of targets in images cannot be fully represented by sparsity. To combine LRSD with characters of target images, a robust principal component analysis combined with top-hat transform (RPCA-THT) is proposed. RPCA-THT optimizes the step of shrinking the sparse component in robust principal component analysis (RPCA). It performs a top-hat transform on the sparse component with a specially designed kernel matrix. Then the sparse component shrinks according to the normalized top-hat transform of the sparse matrix. In this way, targets prefer to be left in the sparse component than clutter. We design a well-performed kernel for distinguishing targets in the sparse matrix. The experimental results on simulation and real data show that the proposed method has better performance than several state-of-the-art clutter suppression algorithms.
Ground penetrating radarclutter suppressionrobust principal component analysistop-hat transformlow-rank sparse decomposition
NETL
NSTL
Taylor & Francis
