Video SAR Moving Target Detection Using Dual Faster R-CNN

Wen, Liwu; Ding, Jian; Loffeld, Otmar

doi:10.1109/jstars.2021.3062176

Cited by 40 publications

(10 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…) Compared with (10) and (11) where the searching scopes are based on the last-frame positioning, the searching scopes in (17) and (18) for shadow or energy are corrected by another whose tracking confidence score is higher. The responses of the corrected searching scopes are then calculated to relocate the target.…”

Section: Tracking Enhancement and Feedbackmentioning

confidence: 99%

“…Compared with conventional tracking methods, the feature-based tracking does not rely on the accurate motion model of the moving target and can work well when faced with low SCNR, especially for the maneuvering targets. A framework using deep neural networks was presented for shadow detection in ViSAR [17], and the target energy was used for the false alarm suppression [18]. The neural-network-based algorithms usually suffer from their limited generalization ability in radar applications.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Video SAR Moving Target Tracking Using Joint Kernelized Correlation Filter

Zhong

Ding

Zhang

2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

Video synthetic aperture radar (ViSAR) has been found very useful for the surveillance of ground moving targets. The target energy can be utilized for ground moving target tracking, while the dynamic shadows of moving targets enable an alternative tracking approach. However, neither of these two approaches can stand alone to provide reliable target tracking. The smeared shadow and energy both degrade the tracking performance when the target is maneuvering.A moving target tracking framework based on the joint kernelized correlation filter (JKCF) has been developed. Based on the feature training of JKCF, the target is tracked by combining its shadow in the sequential SAR imagery and the corresponding energy in the range-Doppler (RD) spectra. Aiming at the problems of tracking drift and collapse, interactive processing is adopted to enhance the target positioning and feature update based on the confidence assessment. By cooperating with the initialization and feature update strategy, the tracking success rate and precision can be improved significantly.

show abstract

Section: Tracking Enhancement and Feedbackmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Video SAR Moving Target Tracking Using Joint Kernelized Correlation Filter

Zhong

Ding

Zhang

2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

show abstract

“…The authors of [18] presented a method of local feature analysis based on single-frame images that can accurately detect the shadow of a moving target. In [19], a dual fast region-based convolutional neural network for SAR images in the range Doppler (RD) spectrum domain was proposed to detect a moving target on the ground, but these algorithms have high requirements for training data [20] and insufficient generalization ability. All these algorithms can extract the shadow of a moving target on the ground.…”

Section: Introductionmentioning

confidence: 99%

Detecting Moving Target on Ground Based on Its Shadow by Using VideoSAR

Chen

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

Video synthetic aperture radar (VideoSAR) can detect and identify a moving target based on its shadow. A slowly moving target has a shadow with distinct features, but it cannot be detected by state-of-the-art difference-based algorithms because of minor variations between adjacent frames. Furthermore, the detection boxes generated by difference-based algorithms often contain such defects as misalignments and fracture. In light of these problems, this study proposed a robust moving target detection (MTD) algorithm for objects on the ground by fusing the background frame detection results and the difference between frames over multiple intervals. We also discuss defects that occur in conventional MTD algorithms. The difference in background frame was introduced to overcome the shortcomings of difference-based algorithms and acquire the shadow regions of objects. This was fused with the multi-interval frame difference to simultaneously extract the moving target at different velocities while identifying false alarms. The results of experiments on empirically acquired VideoSAR data verified the performance of the proposed algorithm in terms of detecting a moving target on the ground based on its shadow.

show abstract

“…However, the features used by these traditional methods are usually simple, which leads to the problem of the background being similar to the shadow, meaning it cannot be easily distinguished. Deep learning methods then emerged to solve shadow tracking due to their high accuracy and fast speed advantages [12][13][14][15][16]. Ding et al [12] presented a framework for shadow-aided moving target detection using deep neural networks, which applied a faster region-based convolutional neural network (Faster-RCNN) [13] to detect shadows in a single frame and used a bi-directional long short-term memory (Bi-LSTM) [14] network to track the shadows.…”

Section: Introductionmentioning

confidence: 99%

“…Zhou et al [15] proposed a framework by combining a modified real-time recurrent regression network and a newly designed trajectory smoothing long short-term memory network to track shadows. Wen et al [16] proposed a moving target tracking method based on the dual Faster-RCNN, which combined the shadow detection results in SAR images and the range-Doppler (RD) spectrum to suppress false alarms for moving target tracking in Video-SAR.…”

Section: Introductionmentioning

confidence: 99%

A Novel Guided Anchor Siamese Network for Arbitrary Target-of-Interest Tracking in Video-SAR

et al. 2021

View full text Add to dashboard Cite

Video synthetic aperture radar (Video-SAR) allows continuous and intuitive observation and is widely used for radar moving target tracking. The shadow of a moving target has the characteristics of stable scattering and no location shift, making moving target tracking using shadows a hot topic. However, the existing techniques mainly rely on the appearance of targets, which is impractical and costly, especially for tracking targets of interest (TOIs) with high diversity and arbitrariness. Therefore, to solve this problem, we propose a novel guided anchor Siamese network (GASN) dedicated to arbitrary TOI tracking in Video-SAR. First, GASN searches for matching areas in the subsequent frames with the initial area of the TOI in the first frame are conducted, returning the most similar area using a matching function, which is learned from general training without TOI-related data. With the learned matching function, GASN can be used to track arbitrary TOIs. Moreover, we also constructed a guided anchor subnetwork, referred to as GA-SubNet, which employs the prior information of the first frame and generates sparse anchors of the same shape as the TOIs. The number of unnecessary anchors is therefore reduced to suppress false alarms. Our method was evaluated on simulated and real Video-SAR data. The experimental results demonstrated that GASN outperforms state-of-the-art methods, including two types of traditional tracking methods (MOSSE and KCF) and two types of modern deep learning techniques (Siamese-FC and Siamese-RPN). We also conducted an ablation experiment to demonstrate the effectiveness of GA-SubNet.

show abstract

Video SAR Moving Target Detection Using Dual Faster R-CNN

Cited by 40 publications

References 26 publications

Video SAR Moving Target Tracking Using Joint Kernelized Correlation Filter

Video SAR Moving Target Tracking Using Joint Kernelized Correlation Filter

Detecting Moving Target on Ground Based on Its Shadow by Using VideoSAR

A Novel Guided Anchor Siamese Network for Arbitrary Target-of-Interest Tracking in Video-SAR

Contact Info

Product

Resources

About