A Compact Methodology to Understand, Evaluate, and Predict the Performance of Automatic Target Recognition

Li, Yanpeng; Li, Xiang; Wang, Hongqiang; Zhuang, Zhihao; Cheng, Yongqiang; Deng, Bin; Wang, Liandong; Zeng, Yonghu; Gao, Lei

doi:10.3390/s140711308

Cited by 13 publications

(12 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Automatic target recognition (ATR) has been applied for many fields, such as social security, environmental monitoring national defence etc. [1–3]. Synthetic aperture radar (SAR) ATR plays an important role in military and civil applications, as SAR can work in all‐weather environments [4, 5].…”

Section: Introductionmentioning

confidence: 99%

Target recognition in synthetic aperture radar images via non‐negative matrix factorisation

Cui

Cao

Yang

et al. 2015

IET Radar, Sonar & Navigation

128

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Target recognition in synthetic aperture radar images via non‐negative matrix factorisation

Cui

Cao

Yang

et al. 2015

IET Radar, Sonar & Navigation

128

View full text Add to dashboard Cite

“…The existing two performance evaluation approaches for target detection cannot offer detailed performance information in this complicated scenario. With the knowledge of performance evaluation for automatic target recognition systems [19] and previous work on radar signal processing [20,21], we developed some evaluation indexes that may fulfill the performance evaluation demand of target detection with an NSVBR. The objectiveness in each section and the data are organized as follows:

We first review the background of an NSVBR and the existing methods in the performance evaluation of target detection with a radar.

…”

Section: Introductionmentioning

confidence: 99%

Performance Evaluation of Target Detection with a Near-Space Vehicle-Borne Radar in Blackout Condition

Wang

et al. 2016

Sensors

Self Cite

View full text Add to dashboard Cite

Radar is a very important sensor in surveillance applications. Near-space vehicle-borne radar (NSVBR) is a novel installation of a radar system, which offers many benefits, like being highly suited to the remote sensing of extremely large areas, having a rapidly deployable capability and having low vulnerability to electronic countermeasures. Unfortunately, a target detection challenge arises because of complicated scenarios, such as nuclear blackout, rain attenuation, etc. In these cases, extra care is needed to evaluate the detection performance in blackout situations, since this a classical problem along with the application of an NSVBR. However, the existing evaluation measures are the probability of detection and the receiver operating curve (ROC), which cannot offer detailed information in such a complicated application. This work focuses on such requirements. We first investigate the effect of blackout on an electromagnetic wave. Performance evaluation indexes are then built: three evaluation indexes on the detection capability and two evaluation indexes on the robustness of the detection process. Simulation results show that the proposed measure will offer information on the detailed performance of detection. These measures are therefore very useful in detecting the target of interest in a remote sensing system and are helpful for both the NSVBR designers and users.

show abstract

“…Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR) plays an important role in military and civil applications, such as social security, environmental monitoring, and national defense [1][2][3][4][5]. Most current researches focus on the pattern features [6][7][8] or pattern classifiers [9,10].…”

Section: Introductionmentioning

confidence: 99%

Hierarchical Recognition System for Target Recognition from Sparse Representations

Cui

Cao

Yang

et al. 2015

Mathematical Problems in Engineering

View full text Add to dashboard Cite

A hierarchical recognition system (HRS) based on constrained Deep Belief Network (DBN) is proposed for SAR Automatic Target Recognition (SAR ATR). As a classical Deep Learning method, DBN has shown great performance on data reconstruction, big data mining, and classification. However, few works have been carried out to solve small data problems (like SAR ATR) by Deep Learning method. In HRS, the deep structure and pattern classifier are combined to solve small data classification problems. After building the DBN with multiple Restricted Boltzmann Machines (RBMs), hierarchical features can be obtained, and then they are fed to classifier directly. To obtain more natural sparse feature representation, the Constrained RBM (CRBM) is proposed with solving a generalized optimization problem. Three RBM variants,L1-RNM,L2-RBM, andL1/2-RBM, are presented and introduced to HRS in this paper. The experiments on MSTAR public dataset show that the performance of the proposed HRS with CRBM outperforms current pattern recognition methods in SAR ATR, like PCA + SVM, LDA + SVM, and NMF + SVM.

show abstract

A Compact Methodology to Understand, Evaluate, and Predict the Performance of Automatic Target Recognition

Cited by 13 publications

References 52 publications

Target recognition in synthetic aperture radar images via non‐negative matrix factorisation

Target recognition in synthetic aperture radar images via non‐negative matrix factorisation

Performance Evaluation of Target Detection with a Near-Space Vehicle-Borne Radar in Blackout Condition

Hierarchical Recognition System for Target Recognition from Sparse Representations

Contact Info

Product

Resources

About