Search

In synthetic aperture radar (SAR) automatic target recognition, it is expensive and time-consuming to annotate the targets. Thus, training a network with a few labeled data and plenty of unlabeled data attracts attention of many researchers. In this article, we design a semisupervised learning framework including self-consistent augmentation rule, mixup-based mixture, and weighted loss, which allows a classification network to utilize unlabeled data during training and ultimately alleviates the demand of labeled data. The proposed self-consistent augmentation rule forces the samples before and after augmentation to share the same labels to utilize the unlabeled data, which can ensure the prominent effect of supervised learning part of the framework for training by balancing amounts of labeled and unlabeled samples in a minibatch, and makes the network achieve better performance. Then, a mixture method is introduced to mix the labeled, unlabeled, and augmented samples for the better involvement of label information in the mixed samples. By using cross-entropy loss for the mixed-labeled mixtures and mean-squared error loss for the mixed-unlabeled mixtures, the total loss is defined as the weighted sum of them. The experiments on the MSTAR data set and OpenSARShip data set show that the performance of the method is not only far better than the state of the art among current semisupervised-based classifiers but also near to the state of the art among the supervised learning-based networks.

Motivated by the intrinsic dynamics of physical motion as well as establishment of target motion model, this article addresses the problem of human motion recognition with ultra wide band (UWB) through-the-wall radar (TWR) in a novel view of range profile serialization. Specifically, we first convert the original radar echoes into range profiles. Then, an auto-encoder network (AEN) with three dense layers is adopted to reduce the dimension and extract the features of each range profile. After that, a gated recurrent unit (GRU) network with two hidden layers is employed to deal with the features of each time-range slice and output the recognition results at each slice in real time. Finally, experimental data with respect to four different behind-wall human motions is collected by self-developed UWB TWR to validate the effectiveness of the proposed model. The results show that the proposed model can validly recognize the human motion serialization and achieve 93% recognition accuracy within the initial 20% duration of the activities (the average durations are 4s, 5.5s, 3s and 4.5s), which is of great significance for real-time human motion recognition.

In the problem of geospatial object detection, the quality and amount of reference boxes significantly impact the detection performance and prediction speed of object detection networks. Nowadays, most of the popular detection methods adopt the anchor mechanism to generate reference boxes. This paper proposed an anchor-free and sliding-window-free deconvolutional region proposal network and constructed a two-stage deconvolutional object detection network. Instead of using an anchor mechanism, we proposed to use a deconvolutional neural network followed by a connected region generation module to generate reference boxes. The comparison experiments and quantitative analysis with NWPU VHR-10 dataset demonstrate that DeRPN can vastly reduce the number of reference boxes and improve the precision of the reference box coordinates. The experiments also suggest that our proposed two-stage object detection network can not only obtain the nearly state-of-the-art detection results but also achieve the prediction speed close to that of the one-stage detection network.

Deep neural networks have been widely applied and researched in remote sensing image scene classification and achieved a great success. However, deep supervised network heavily relies on a large amount of labeled data. The annotation is difficult and time-consuming to obtain but the unlabeled data are comparably easier to get. Considering that, we introduce a semi-supervised learning framework for remote sensing image scene classification. The network is trained by a novel adaptive perturbation training method. The experiments on NWPU-RESISC45 dataset prove that the introduced semi-supervised classification method can achieve higher classification accuracy with unlabeled data compared with the corresponding supervised classifier, and the designed adaptive perturbation training can further improve the performance of the semi-supervised learning-based classification network.

Stable and efficient ground moving target tracking and refocusing is a hard task in synthetic aperture radar (SAR) data processing. Since shadows in video-SAR indicate the actual positions of moving targets at different moments without any displacement, shadow-based methods provide a new approach for ground moving target processing. This paper constructs a novel framework to refocus ground moving targets by using shadows in video-SAR. To this end, an automatic-registered SAR video is first obtained using the video-SAR back-projection (v-BP) algorithm. The shadows of multiple moving targets are then tracked using a learning-based tracker, and the moving targets are ultimately refocused via a proposed moving target back-projection (m-BP) algorithm. With this framework, we can perform detecting, tracking, imaging for multiple moving targets integratedly, which significantly improves the ability of moving-target surveillance for SAR systems. Furthermore, a detailed explanation of the shadow of a moving target is presented herein. We find that the shadow of ground moving targets is affected by a target’s size, radar pitch angle, carrier frequency, synthetic aperture time, etc. With an elaborate system design, we can obtain a clear shadow of moving targets even in X or C band. By numerical experiments, we find that a deep network, such as SiamFc, can easily track shadows and precisely estimate the trajectories that meet the accuracy requirement of the trajectories for m-BP.

SAR automatic object classification plays an important role in many SAR-based applications. Currently, the research about SAR automatic object classification focuses on supervised learning, which needs a large amount of annotated data for training the network. We find that clustering can achieve object classification based on unsupervised learning. Considering time consumption and expensive cost of manual labeling, unsupervised learning method has a prominent advantage since it does not need manually annotated data. In this paper, we introduce an unsupervised learning method to achieve SAR object classification with no labeled data. The introduced method models object clustering as a recurrent process, in which data samples are gradually clustered together according to their similarity, and feature representation of them is obtained simultaneously. The experiments on MSTAR dataset and quantitative analysis demonstrate the effectiveness of the introduced method.

Deep learning networks are widely being applied to remote sensing image recognition and have achieved promising results. In this paper, we researched the influence of background with different scattering characteristics for synthetic aperture radar (SAR) target recognition based on convolutional neural network (CNN). Firstly, a two-parameter CFAR image segmentation method based on Weibull distribution was used to extracted SAR target and its shadow. And then, SAR datasets with road, farmland and grassland background environment is synthesized to analyze the CNN classifier. Experiments results show that the method by mixing training sets with different background together can improve the recognization rate when the backgrounds are complex.

Conventional synthetic aperture radar (SAR) reconstructs the illuminated scene via fast Fourier transform (FFT), which results in high sidelobe level, and poor cross-range resolution due to the finite synthetic aperture length. In this paper, we formulate a sparse reconstruction method for SAR imaging based on the covariance fitting criterion, which assumes that only a few strong scatters exist in the whole scene. The method is able to fully control over the sparsity level and reconstruct the scenario in an adaptive manner. Experimental results with real SAR data show the better performance of our method compared with the conventional methods in terms of resolution improvement and sidelobe suppression.

Due to finite of information dimension of single bistatic synthetic aperture radar (SAR), to expand more information about ground objects we research multistatic SAR which can be decomposed into groups of bistatic SAR. It is known that scattering properties of the different viewing angles is different. In this paper, we focus on system of single transmitter and triple receivers and use polar format algorithm(PFA) to obtain images of ground objects for the triple receivers. A geometric distortion correction method is proposed due to elevation of ground objects. After the distortion correction, the three SAR images are registrated, then we put images into red, green, blue(RGB) channels respectively to realize fake color synthesis, and thus realize information fusion.

For the existence of speckles, many standard optical image processing methods, such as classification, segmentation, and registration, are restricted to synthetic aperture radar (SAR) images. In this work, an end-to-end deep multi-scale recurrent network (MSR-net) for SAR image despeckling is proposed. The multi-scale recurrent and weights sharing strategies are introduced to increase network capacity without multiplying the number of weights parameters. A convolutional long short-term memory (convLSTM) unit is embedded to capture useful information and helps with despeckling across scales. Meanwhile, the sub-pixel unit is utilized to improve the network efficiency. Besides, two criteria, edge feature keep ratio (EFKR) and feature point keep ratio (FPKR), are proposed to evaluate the performance of despeckling capacity for SAR, which can assess the retention ability of the despeckling algorithm to edge and feature information more effectively. Experimental results show that our proposed network can remove speckle noise while preserving the edge and texture information of images with low computational costs, especially in the low signal noise ratio scenarios. The peak signal to noise ratio (PSNR) of MSR-net can outperform traditional despeckling methods SAR-BM3D (Block-Matching and 3D filtering) by more than 2 dB for the simulated image. Furthermore, the adaptability of optical image processing methods to real SAR images can be enhanced after despeckling.