Baiyuan Ding

A robust similarity measure for attributed scattering center sets with application to SAR ATR

This paper proposes a robust similarity measure for two attributed scattering center (ASC) sets and applies it to synthetic aperture radar (SAR) automatic target recognition (ATR). The extraction uncertainty of an individual ASC is modeled by an adaptive Gaussian distribution according to its attributes. Then the distance between two individual ASCs is defined as the Kullback-Leibler (KL) divergence between two Gaussian distributions which model the uncertainties of those two ASCs. The proposed distance measure can better exploit the inner discrepancy between individual ASCs compared with the Euclid distance or Mahalanobis distance. Based on the proposed distance measure, a cost matrix which contains the costs of false and missing ASCs is built and the Hungarian algorithm is employed to build a one-to-one correspondence between two ASC sets. A threshold method is carried out to further evaluate the Hungarian assignment. Afterwards, a robust similarity measure is designed to evaluate the similarity between the two ASC sets which comprehensively considers the influences of the missing and false ASCs as well as the disproportionate contributions by different ASCs. Finally, the target type is determined by the similarities between the testing image and various types of template targets. Experimental results on the moving and stationary target acquisition and recognition (MSTAR) dataset verify the validity and robustness of the proposed method.

Robust method for the matching of attributed scattering centers with application to synthetic aperture radar automatic target recognition

Abstract. This paper proposes a robust method for the matching of attributed scattering centers (ASCs) with application to synthetic aperture radar automatic target recognition (ATR). For the testing image to be classified, ASCs are extracted to match with the ones predicted by templates. First, Hungarian algorithm is employed to match those two ASC sets initially. Then, a precise matching is carried out through a threshold method. Point similarity and structure similarity are calculated, which are fused to evaluate the overall similarity of the two ASC sets based on the Dempster–Shafer theory of evidence. Finally, the target type is determined by such similarities between the testing image and various types of targets. Experiments on the moving and stationary target acquisition and recognition data verify the validity of the proposed method.

Target Recognition in Synthetic Aperture Radar Images via Matching of Attributed Scattering Centers

This paper presents an approach for attributed scattering center (ASC) matching with application to synthetic aperture radar (SAR) automatic target recognition (ATR). A statistics-based distance measure is designed to evaluate the distance between individual ASCs. Afterwards, the Hungarian algorithm is employed to build a one-to-one correspondence between two ASC sets. Based on the correspondence, a global similarity and a local similarity are designed to comprehensively evaluate the global consistency and structural correlation between those two ASC sets. The two similarities comprehensively exploit the inner correlation between the two ASC sets, thus providing a reliable and robust similarity measure for SAR ATR. The two similarities are then fused based on the Dempster–Shafer evidence theory to determine the target type by the maximum belief rule. Extensive experiments conducted on the moving and stationary target acquisition and recognition dataset and the comparison with several state-of-the-art methods demonstrate the validity and robustness of the proposed method.

Target recognition in SAR images by exploiting the azimuth sensitivity

ABSTRACT Azimuth sensitivity is a significant characteristic of synthetic aperture radar (SAR) images. Most of the previous SAR target recognition algorithms try to cope with the property by pose estimation or training classifiers which are not sensitive to azimuth. Actually, the azimuth sensitivity can provide discriminative information for target recognition as a supplement to the original spatial image (SI). This letter describes the azimuth sensitivity by the azimuth sensitivity image (ASI) which is constructed by comparing the sub-aperture images of the SI. Then the SI and ASI are classified by the sparse representation-based classification (SRC), respectively. Afterwards, a score-level fusion is employed to combine the two results for robust target recognition. Extensive experiments are conducted on the moving and stationary target acquisition and recognition (MSTAR) dataset and the performance is compared with several state-of-the-art methods. The experimental results show that the ASI can complement the SI for effective and robust target recognition.

Data Augmentation by Multilevel Reconstruction Using Attributed Scattering Center for SAR Target Recognition

The quality of synthetic aperture radar (SAR) images and the completeness of the template database are two important factors in template-based SAR automatic target recognition. This letter gives a solution to the two factors by multilevel reconstruction of SAR targets using attributed scattering centers (ASCs). The ASCs of original SAR images are extracted to reconstruct the target’s image, which not only reduces the noise and background clutters but also keeps the electromagnetic characteristics of the target. Template database are reconstructed at multilevels to simulate various extents of ASC absence in the extended operation conditions. Therefore, the quality of SAR images as well as the completeness of the template database is augmented. Features are extracted from the augmented SAR images, and the classifier is trained by the augmented database for target recognition. Experimental results on the moving and stationary target acquisition and recognition data set demonstrate the validity of the proposed method.

Target recognition of SAR images based on multi-resolution representation

ABSTRACT This letter presents a target recognition method for synthetic aperture radar (SAR) images by exploiting the multi-resolution information. For the training samples, images with lower resolutions are generated to construct a multi-resolution dictionary for the sparse representation-based classification (SRC). Then the test sample will be classified based on the augmented dictionary. The multi-resolution representation of the training samples can not only promote the representation capability of the dictionary but also enhance the robustness to the resolution variance of the test sample due to the sensor variation. Experiments are conducted on the moving and stationary target acquisition and recognition (MSTAR) dataset to demonstrate the validity of the proposed method.

CFAR Detection of Moving Range-Spread Target in White Gaussian Noise Using Waveform Contrast

In wideband stepped-frequency radar systems, relative motions between radar and target can induce high distortions of high-resolution range profiles (HRRPs), such as range migration, shape deformation, and signal-to-noise-ratio (SNR) loss. These distortions, if ignored, can lead to unacceptable performance deterioration in detection. To solve this problem, a new algorithm for detecting moving range-spread targets (RSTs) is proposed in this letter. The proposed detector utilizes the waveform contrast of the HRRP to perform both motion compensation and constant false-alarm rate target detection, and it is simple and robust even for low-SNR scenarios. Simulated experiments are carried out to verify the effectiveness and advantages of the proposed detector.

Evaluation of target segmentation on SAR target recognition

Target segmentation of synthetic aperture radar (SAR) images is one of the challenging problems in SAR image interpretation, which often serves as a processing step for SAR target recognition. Target segmentation tries to separate the target from the background thus eliminating the interference of background noises or clutters. However, the segmentation may also discard a part of the target characteristics and the target shadow which also contain discriminative information for target recognition. Then the tradeoff between interference elimination and discriminability loss will cause some effects on the target recognition. This paper aims to evaluate the influence of target segmentation on target recognition. Target recognition under standard operating condition (SOC) and several extended operating conditions (EOCs), i.e., depression angle variance and noise corruption, is conducted on the moving and stationary target acquisition and recognition (MSTAR) dataset using the original image and segmented target image, respectively. Moreover, the recognition performance under target segmentation errors is also evaluated. By comparing the recognition performance, the effects of target segmentation can be illustrated.

Three dimensional electromagnetic model guided scattering center extraction

Scattering center extraction from Synthetic Aperture Radar (SAR) data is a critical step in model-based SAR automatic target recognition (ATR). Three dimensional electromagnetic model (3D em-model) provides a concise and physically relevant description of targets electromagnetic scattering behavior by a set of representative scattering centers. In this paper, 3D em-model is used to guide scattering center extraction in SAR data. Firstly, 3D em-model is projected to the 2D measurement plane to predict the scattering region and the attributed parameters of each scattering center. Then, scattering center is extracted in the corresponding region predicted by 3D em-model with the model predicted parameters as an initial guess. Finally, a search strategy is adopted to optimize the scattering region and find the optimal parameters. Experiments using data simulated by a high-frequency electromagnetic code verify the validity of this method.

Target recognition in SAR images via Gaussian mixture modeling of attributed scattering center set

Attributed scattering center (ASC) is an important feature for synthetic aperture radar (SAR) automatic target recognition (ATR). This paper uses Gaussian mixture model (GMM) to model the uncertainties of two ASC sets which are predicted by the template image and extracted from the testing image respectively. Then the distance between the two ASC sets is measured by the L2 distance between their GMMs. Finally, the target type is determined by the distances between the extracted ASC set and various types of predicted ASC sets using a nearest neighbor (NN) classifier. The proposed method avoids the problem of building a one-to-one correspondence between ASC sets so it is efficient and insensitive to noise-caused error and partial occlusion. Experiments on the moving and stationary acquisition and recognition (MSTAR) dataset demonstrate the validity and efficiency of the proposed method.