The algorithm of synthetic aperture radar automatic target recognition (SAR-ATR) is generally composed of the extraction of a set of features that transform the raw input into a representation, followed by a trainable classifier. The feature extractor is often hand designed with domain knowledge and can significantly impact the classification accuracy. By automatically learning hierarchies of features from massive training data, deep convolutional networks (ConvNets) recently have obtained state-of-the-art results in many computer vision and speech recognition tasks. However, when ConvNets was directly applied to SAR-ATR, it yielded severe overfitting due to limited training images. To reduce the number of free parameters, we present a new all-convolutional networks (A-ConvNets), which only consists of sparsely connected layers, without fully connected layers being used. Experimental results on the Moving and Stationary Target Acquisition and Recognition (MSTAR) benchmark data set illustrate that A-ConvNets can achieve an average accuracy of 99% on classification of ten-class targets and is significantly superior to the traditional ConvNets on the classification of target configuration and version variants.

Target Classification Using the Deep Convolutional Networks for SAR Images

Many methods have been proposed to improve the performance of synthetic aperture radar (SAR) target recognition but seldom consider the issues in real-world recognition systems, such as the invariance under target translation, the invariance under speckle variation in different observations, and the tolerance of pose missing in training data. In this letter, we investigate the capability of a deep convolutional neural network (CNN) combined with three types of data augmentation operations in SAR target recognition. Experimental results demonstrate the effectiveness and efficiency of the proposed method. The best performance is obtained by using the CNN trained by all types of augmentation operations, showing that it is a practical approach for target recognition in challenging conditions of target translation, random speckle noise, and missing pose.

Convolutional Neural Network With Data Augmentation for SAR Target Recognition

Tremendous progress has been made in object recognition with deep convolutional neural networks (CNNs), thanks to the availability of large-scale annotated dataset. With the ability of learning highly hierarchical image feature extractors, deep CNNs are also expected to solve the Synthetic Aperture Radar (SAR) target classification problems. However, the limited labeled SAR target data becomes a handicap to train a deep CNN. To solve this problem, we propose a transfer learning based method, making knowledge learned from sufficient unlabeled SAR scene images transferrable to labeled SAR target data. We design an assembled CNN architecture consisting of a classification pathway and a reconstruction pathway, together with a feedback bypass additionally. Instead of training a deep network with limited dataset from scratch, a large number of unlabeled SAR scene images are used to train the reconstruction pathway with stacked convolutional auto-encoders (SCAE) at first. Then, these pre-trained convolutional layers are reused to transfer knowledge to SAR target classification tasks, with feedback bypass introducing the reconstruction loss simultaneously. The experimental results demonstrate that transfer learning leads to a better performance in the case of scarce labeled training data and the additional feedback bypass with reconstruction loss helps to boost the capability of classification pathway.

Transfer Learning with Deep Convolutional Neural Network for SAR Target Classification with Limited Labeled Data

An improved Otsu method using the weighted object variance for defect detection

In this paper, the classification via sprepresentation and multitask learning is presented for target recognition in SAR image. To capture the characteristics of SAR image, a multidimensional generalization of the analytic signal, namely the monogenic signal, is employed. The original signal can be then orthogonally decomposed into three components: 1) local amplitude; 2) local phase; and 3) local orientation. Since the components represent the different kinds of information, it is beneficial by jointly considering them in a unifying framework. However, these components are infeasible to be directly utilized due to the high dimension and redundancy. To solve the problem, an intuitive idea is to define an augmented feature vector by concatenating the components. This strategy usually produces some information loss. To cover the shortage, this paper considers three components into different learning tasks, in which some common information can be shared. Specifically, the component-specific feature descriptor for each monogenic component is produced first. Inspired by the recent success of multitask learning, the resulting features are then fed into a joint sparse representation model to exploit the intercorrelation among multiple tasks. The inference is reached in terms of the total reconstruction error accumulated from all tasks. The novelty of this paper includes 1) the development of three component-specific feature descriptors; 2) the introduction of multitask learning into sparse representation model; 3) the numerical implementation of proposed method; and 4) extensive comparative experimental studies on MSTAR SAR dataset, including target recognition under standard operating conditions, as well as extended operating conditions, and the capability of outliers rejection.

SAR Target Recognition via Joint Sparse Representation of Monogenic Signal

In this letter, the classification via sparse representation of the monogenic signal is presented for target recognition in SAR images. To characterize SAR images, which have broad spectral information yet spatial localization, the monogenic signal is performed. Then an augmented monogenic feature vector is generated via uniform down-sampling, normalization and concatenation of the monogenic components. The resulting feature vector is fed into a recently developed framework, i.e., sparse representation based classification (SRC). Specifically, the feature vectors of the training samples are utilized as the basis vectors to code the feature vector of the test sample as a sparse linear combination of them. The representation is obtained via ${\ell _1}$ -norm minimization, and the inference is reached according to the characteristics of the representation on reconstruction. Extensive experiments on MSTAR database demonstrate that the proposed method is robust towards noise corruption, as well as configuration and depression variations.

Sparse Representation of Monogenic Signal: With Application to Target Recognition in SAR Images

Traditional two-dimensional (2D) Otsu method supposes that the sum of probabilities of diagonal quadrants in 2D histogram is approximately one. This studies experiments and theory prove that the sum of probabilities of off-diagonal quadrants in 2D histogram is not always very small and this could not be neglected. Therefore the assumption mentioned above in 2D Otsu method is inadequately reasonable. In this study, an improved 2D Otsu segmentation method and recursive algorithm are proposed. By calculating probabilities of diagonal quadrants in 2D histogram separately, modified method is acquired. Experimental results show that proposed method can obtain better performance of segmentation than the traditional 2D Otsu method. The computation complexity of improved 2D Otsu method is equal to traditional 2D Otsu method.

Modified two-dimensional otsu image segmentation algorithm and fast realisation

Automatic target recognition has been widely studied over the years, yet it is still an open problem. The main obstacle consists in extended operating conditions, e.g. ., depression angle change, configuration variation, articulation, and occlusion. To deal with them, this paper proposes a new classification strategy. We develop a new representation model via the steerable wavelet frames. The proposed representation model is entirely viewed as an element on Grassmann manifolds. To achieve target classification, we embed Grassmann manifolds into an implicit reproducing Kernel Hilbert space (RKHS), where the kernel sparse learning can be applied. Specifically, the mappings of training sample in RKHS are concatenated to form an overcomplete dictionary. It is then used to encode the counterpart of query as a linear combination of its atoms. By designed Grassmann kernel function, it is capable to obtain the sparse representation, from which the inference can be reached. The novelty of this paper comes from: 1) the development of representation model by the set of directional components of Riesz transform; 2) the quantitative measure of similarity for proposed representation model by Grassmann metric; and 3) the generation of global kernel function by Grassmann kernel. Extensive comparative studies are performed to demonstrate the advantage of proposed strategy.

Classification via Sparse Representation of Steerable Wavelet Frames on Grassmann Manifold: Application to Target Recognition in SAR Image

This letter addresses the polarimetric synthetic aperture radar target detection using the magnitude of the (2, 3) term in the sample averaged coherency matrix. The theoretical analysis demonstrates that such term reveals the difference between the nonreflection symmetric targets and natural clutters. The statistical models for such term are derived within different degrees of homogeneity. Based on the statistical models, an automatic constant-false-alarm-rate detection scheme is completed. The parameter estimation and the solution for the detection threshold are given in detail. Experimental results demonstrate the capability of the proposed approach for detecting ships, oil stores, buildings, etc., in homogeneous and heterogeneous areas.

Na Wang

Papers

SAR Target Recognition via Joint Sparse Representation of Monogenic Signal

Sparse Representation of Monogenic Signal: With Application to Target Recognition in SAR Images

Modified two-dimensional otsu image segmentation algorithm and fast realisation

Classification via Sparse Representation of Steerable Wavelet Frames on Grassmann Manifold: Application to Target Recognition in SAR Image

Polarimetric SAR Target Detection Using the Reflection Symmetry