The major clutter in a typical ground penetrating radar (GPR) B-scan image is the ground surface clutter. It can severely obscure or distort the subsurface target response, especially when the surface profile variation is large and the target reflection signal is overlapped with the surface clutter. In this paper, we propose a deep auto-encoder based method for reducing the rough surface clutter in GPR images. In our method, the rough surface clutter reduction is formulated as an anomaly detection problem. The rough surface region in a B-scan image is splitted into small patches which are used as the training data set to train a deep auto-encoder. The trained autoencoder captures the patterns of the rough surface patches. After training, the whole B-scan image is divided into small patches of the same size as the training patches and each of which is fed into the auto-encoder to compute an anomaly score. To reduce the clutter, we design a novel method that adopts a weighted sum approach to aggregate all patches based on their anomaly scores. For performance evaluation, simulation and laboratory tests are conducted, and the clutter-to-rest ratios (CRR) are computed to quantify the clutter reduction effectiveness. The CRR of the simulated B-scan image is reduced from 4.24 dB to -14.06 dB, and the CRRs of two measured B-scans are reduced from 3.71 dB to -20.45 dB, and from 8.62 dB to -8.07 dB, respectively.
In this paper, we present an experimental evaluation of recently proposed Supervised Reciprocal Filter approaches for the compression of OFDM-radar signals. The range-Doppler map is usually evaluated using a suboptimal batches algorithm, after fragmenting the signal in batches with length equal to the OFDM symbol. Using “OFDM fragmentation” requires symbol synchronization and sets constraints on the non-ambiguous Range-Doppler area of targets that can be detected with limited Signal-to-Noise Ratio (SNR) loss. Supervised Reciprocal Filters have been recently proposed to operate with batches of longer lengths than the OFDM symbol without requiring any synchronization. In this paper we extend the study to include the case of batches equal to a fraction of the OFDM symbol, which provides higher flexibility to adapt the processor to the range-Doppler scenario of interest. These filters have been shown to contain the large SNR losses obtained with a direct application of the Reciprocal Filter (RF) with the non-OFDM fragmentation. Moreover, they have been shown theoretically to preserve the benefits of the RF over the Matched Filter (MF) against the clutter-limited scenarios. To assess the performance of the Supervised Filter against a real scenario, an acquisition campaign has been carried out using the Sapienza experimental passive radar along the coast north of Rome, against a maritime traffic scenario, including non-cooperative vessels, as well as a cooperating small boat equipped with differential GPS positioning registration tools. The effectiveness of the proposed approaches is validated by applying them to experimental data from a PBR application exploiting DVB-T transmissions.
Ship detection in synthetic aperture radar (SAR) images is a hot pot in the remote sensing (RS) field. However, most existing deep learning (DL)-based methods only focus on the single-polarization SAR ship detection without leveraging the rich dual-polarization SAR features, which poses a huge obstacle to the further model performance improvement. One problem for solution is how to fully excavate polarization characteristics using a convolution neural network (CNN). To address the above problem, we propose a novel group-wise feature fusion R-CNN (GWFF R-CNN) for dual-polarization SAR ship detection. Different from raw Faster R-CNN, GWFF R-CNN embeds a group-wise feature fusion module (GWFF module) into the subnetwork of Faster R-CNN, which enables group-wise feature fusion between polarization features and multi-scale ship features. Finally, the experiments on the dual-polarization SAR ship detection dataset (DSSDD) demonstrate that GWFF R-CNN can yield a ~4.1 F1 improvement and a ~2.9 average precision (AP) improvement, compared with Faster R-CNN.
This paper proposes a novel passive nonlinear tag design, fabrication, and tests using a second-order harmonic system with various antenna configurations. The tags' antennas were fabricated with laser-cut pattern masks and conductive materials, and their performance was greatly enhanced by the multi-element design and Schottky diode's nonlinearity. The detectable distance of up to 6.7 ft of the tags was tested using a benchtop setup of a 7.9 GHz transmitter and a 15.8 GHz receiver antenna without any signal amplification. The lightweight and miniature tags have great potential for extensive applications in locating objects.
When radar detects a high-speed maneuvering target, not only will the phenomena of range migration (RM) and Doppler migration (DM) appear, but also the phenomenon of range ambiguity, which poses challenges to the traditional accumulation processing method. In this paper, we first establish the target echo model with range ambiguity based on the spatial geometric model. On this basis, we propose a coherent integration method based on the modulo generalized Radon Fourier transform (MGRFT). By performing the modulo addressing operation during the joint search of motion parameters, the proposed method can correct RM and DM and deal with the problem of trajectory breakage under range ambiguity so as to achieve the coherent integration of echo energy and effectively improve the signal-to-noise ratio (SNR). Finally, experimental results demonstrate the effectiveness of the proposed algorithm.
