Sparse Representation-Based SAR Image Target Classification on the 10-Class MSTAR Data Set

Abstract

Recent years have witnessed an ever-mounting interest in the research of sparse representation. The framework, Sparse Representation-based Classification (SRC), has been widely applied as a classifier in numerous domains, among which Synthetic Aperture Radar (SAR) target recognition is really challenging because it still is an open problem to interpreting the SAR image. In this paper, SRC is utilized to classify a 10-class moving and stationary target acquisition and recognition (MSTAR) target, which is a standard SAR data set. Before the classification, the sizes of the images need to be normalized to maintain the useful information, target and shadow, and to suppress the speckle noise. Specifically, a preprocessing method is recommended to extract the feature vectors of the image, and the feature vectors of the test samples can be represented by the sparse linear combination of basis vectors generated by the feature vectors of the training samples. Then the sparse representation is solved by l 1 -norm minimization. Finally, the identities of the test samples are inferred by the reconstructive errors calculated through the sparse coefficient. Experimental results demonstrate the good performance of SRC. Additionally, the average recognition rate under different feature spaces and the recognition rate of each target are discussed.

Robust Sonar ATR Through Bayesian Pose-Corrected Sparse Classification

“…Note that each image was normalized). some considerations made in SAR ATR and echocardiogram image classification [10], [48]. For the tests whose mean recall and precision statistics can be found in Figures 11 and 12, respectively, we constructed six models for each algorithm off of forty different training images per class and tested them against thirty six Sonar images (nine per class).…”

Target Recognition in SAR Images Based on Information-Decoupled Representation

“…Because of the non-convex 0 -norm objective, the optimization problem in Equation (2) is an NP-hard problem. Typical approaches to solve the problem are either approximating the original problem with 1 -norm based convex relaxation such as an 1 -minimization [38] or resorting to greedy schemes such as orthogonal matching pursuit (OMP) [21,34,39,40]. The detailed implementation of the OMP algorithm to solve Equation (2) is presented in Algorithm 1 [46], which will be used in the following target recognition.…”

“…After the sparse coefficient vectorα is solved, the SRC decides the identity of the test sample as the class with the minimum reconstruction error [21,34,39,40].…”

“…As indicated in [47], the random projection is a good "partner" for "compressive classification" including SRC. To solve the sparse coefficients, the OMP algorithm is used, which is demonstrated to be effective in SAR ATR [21,34,39,40]. …”

“…With the fast development of pattern recognition and machine learning techniques, many advanced classifiers [30][31][32][33][34][35][36][37] have been successfully applied to SAR ATR, such as the support vector machine (SVM) [28,29], sparse representation-based classification (SRC) [21,[31][32][33], convolutional neural network (CNN) [35], adaptive boosting (Adaboost) [36] and discriminative graphical models [37]. Among these classifiers, SRC is notably robust to EOCs such as noise corruption and partial occlusion [38].…”