Multi-Block Mixed Sample Semi-Supervised Learning for SAR Target Recognition

Ye Tian,Liguo Zhang,Jianguo Sun,Pengyuan Qi,Guisheng Yin

doi:10.3390/rs13030361

Abstract

In recent years, synthetic aperture radar (SAR) automatic target recognition has played a crucial role in multiple fields and has received widespread attention. Compared with optical image recognition with massive annotation data, lacking sufficient labeled images limits the performance of the SAR automatic target recognition (ATR) method based on deep learning. It is expensive and time-consuming to annotate the targets for SAR images, while it is difficult for unsupervised SAR target recognition to meet the actual needs. In this situation, we propose a semi-supervised sample mixing method for SAR target recognition, named multi-block mixed (MBM), which can effectively utilize the unlabeled samples. During the data preprocessing stage, a multi-block mixed method is used to interpolate a small part of the training image to generate new samples. Then, the new samples are used to improve the recognition accuracy of the model. To verify the effectiveness of the proposed method, experiments are carried out on the moving and stationary target acquisition and recognition (MSTAR) data set. The experimental results fully demonstrate that the proposed MBM semi-supervised learning method can effectively address the problem of annotation insufficiency in SAR data sets and can learn valuable information from unlabeled samples, thereby improving the recognition performance.

Highlights

Automatic target recognition (ATR) for synthetic aperture radar (SAR) has been widely applied in mineral resource exploration, geographic information collection, and marine monitoring, due to its all-weather, all-time, and long-range operation and high-resolution imaging superiority ability [1,2,3,4,5,6]
The moving and stationary target acquisition and recognition (MSTAR) data set [28] is a public data set created by the U.S Air Force Laboratory, which consists of SAR images of ten classes of military vehicles with ground targets and is divided into two sub-datasets: a training data set and a testing data set
The training images are obtained at a 17◦ depression angle, and the testing images are captured at a 15◦ depression angle

Summary

Introduction

Automatic target recognition (ATR) for SAR has been widely applied in mineral resource exploration, geographic information collection, and marine monitoring, due to its all-weather, all-time, and long-range operation and high-resolution imaging superiority ability [1,2,3,4,5,6]. Among techniques for ATR are the feature-based methods, which extract features from SAR images to feed into the classifier for recognition [7,8,9,10]. These methods can improve the accuracy of target recognition, and can reduce the requirement of the sample amount. Unlike traditional machine learning methods [15], which require handcrafted features, CNNs can automatically learn effective hierarchical image features to achieve higher recognition accuracy [16]

Objectives

Methods

Results

Conclusion