SAR ATR of Ground Vehicles Based on LM-BN-CNN

Zhou, Feng; Wang, Li; Bai, Xueru; Ye, Hui

doi:10.1109/tgrs.2018.2849967

Cited by 87 publications

(42 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Reference [14] shows that, although the target region has been removed from original MSTAR images, the nearest neighbor classifier still achieves high classification accuracy, proving that the clutter in the training and test images of the MSTAR dataset is highly correlated. Reference [15] also proves that background clutter in the MSTAR dataset will disturb the recognition results of CNN. The target region is segmented out from the original SAR images according to references [15] and [30] to mitigate the impact of background clutter on network training and testing, as shown by Figure 9.…”

Section: Effect Of Clutter and Data Generationmentioning

confidence: 95%

“…Reference [15] also proves that background clutter in the MSTAR dataset will disturb the recognition results of CNN. The target region is segmented out from the original SAR images according to references [15] and [30] to mitigate the impact of background clutter on network training and testing, as shown by Figure 9. The original 128 × 128 image is cropped to 60 × 60 to reduce the computational cost, because the target only occupies a small region at the center of the original image.…”

Section: Effect Of Clutter and Data Generationmentioning

confidence: 95%

“…Figure 3 shows main steps of the training and test stages to give a brief view of the proposed method. Firstly, image segmentation is utilized to remove the background clutter [15,30]. Subsequently, the segmented training images are input into the ESENet to learn weights, and all of the weights in the ESENet are fixed when the training stage ends.…”

Section: Sar Atr Based On Esenetmentioning

confidence: 99%

“…It may be impossible that the target location may change in real world situations, and various background clutters instead of a fixed type should accompany the corresponding SAR image. Therefore, we exclude such correlation by target region segmentation [15] and generate the MSTAR pure target dataset for fair comparison and an evaluation of SAR ATR algorithms.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

SAR ATR of Ground Vehicles Based on ESENet

2019

Self Cite

View full text Add to dashboard Cite

In recent studies, synthetic aperture radar (SAR) automatic target recognition (ATR) algorithms that are based on the convolutional neural network (CNN) have achieved high recognition rates in the moving and stationary target acquisition and recognition (MSTAR) dataset. However, in a SAR ATR task, the feature maps with little information automatically learned by CNN will disturb the classifier. We design a new enhanced squeeze and excitation (enhanced-SE) module to solve this problem, and then propose a new SAR ATR network, i.e., the enhanced squeeze and excitation network (ESENet). When compared to the available CNN structures that are designed for SAR ATR, the ESENet can extract more effective features from SAR images and obtain better generalization performance. In the MSTAR dataset containing pure targets, the proposed method achieves a recognition rate of 97.32% and it exceeds the available CNN-based SAR ATR algorithms. Additionally, it has shown robustness to large depression angle variation, configuration variants, and version variants.

show abstract

Section: Effect Of Clutter and Data Generationmentioning

confidence: 95%

Section: Effect Of Clutter and Data Generationmentioning

confidence: 95%

Section: Sar Atr Based On Esenetmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

SAR ATR of Ground Vehicles Based on ESENet

2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…Corresponding results are shown in Table 2. To make the comparison between A-convnet and the proposed algorithm convenient, we refer to the structure of A-convnet constructed in [26,27]. e structure that results in the best recognition results for the 64 × 64 pixels image is displayed in Figure 2.…”

Section: Target Configuration Recognition On Various Targetsmentioning

confidence: 99%

Product Dictionary Learning-Based SAR Target Configuration Recognition

Liu

Chen

et al. 2020

International Journal of Antennas and Propagation

View full text Add to dashboard Cite

Dictionary construction is a key factor for the sparse representation-(SR-) based algorithms. It has been verified that the learned dictionaries are more effective than the predefined ones. In this paper, we propose a product dictionary learning (PDL) algorithm to achieve synthetic aperture radar (SAR) target configuration recognition. e proposed algorithm obtains the dictionaries from a statistical standpoint to enhance the robustness of the proposed algorithm to noise. And, taking the inevitable multiplicative speckle in SAR images into account, the proposed algorithm employs the product model to describe SAR images. A more accurate description of the SAR image results in higher recognition rates. e accuracy and robustness of the proposed algorithm are validated by the moving and stationary target acquisition and recognition (MSTAR) database.

show abstract