Airports are important targets in both military and civilian tasks. Synthetic aperture radar (SAR)-based airport detection has received increasing attention in recent years. However, due to the high cost of the SAR imaging and annotation process, there is no publicly available SAR dataset for airport detection, which leads to the fact that deep learning methods have not been applied to airport detection and no unified evaluation benchmark for traditional methods. To provide a benchmark for airport detection studies in SAR images, a large SAR airport dataset is presented in this article. It contains 624 SAR images from Sentinel 1B, covering 104 instances of airports with different scales, orientations, and shapes, which can realistically reflect the real world. Experiments with four deep learning-based methods and three traditional methods on this dataset demonstrate its effectiveness and challenge. It serves for the development of state-of-the-art airport detection algorithms or other related tasks. In addition, we found that airport runways in SAR images always have some parallel line segments. Inspired by this, the fusion of deep features and line segments is achieved by designing a line segment detector branch, which further improves the accuracy of airport detection.
In this paper, we propose a fast Line Segment Detection algorithm for Polarimetric synthetic aperture radar (PolSAR) data (PLSD). We introduce the Constant False Alarm Rate (CFAR) edge detector to obtain the gradient map of the PolSAR image, which tests the equality of the covariance matrix using the test statistic in the complex Wishart distribution. A new filter configuration is applied here to save time. Then, the Statistical Region Merging (SRM) framework is utilized for the generation of line-support regions. As one of our main contributions, we propose a new Statistical Region Merging algorithm based on gradient Strength and Direction (SRMSD). It determines the merging predicate with consideration of both gradient strength and gradient direction. For the merging order, we set it by bucket sort based on the gradient strength. Furthermore, the pixels are restricted to belong to a unique region, making the algorithm linear in time cost. Finally, based on Markov chains and a contrario approach, the false alarm control of line segments is implemented. Moreover, a large scene airport detection method is designed based on the proposed line segment detection algorithm and scattering characteristics. The effectiveness and applicability of the two methods are demonstrated with PolSAR data provided by UAVSAR.
Radar, as a contactless, nonintrusive, weather- and light-independent sensor, works in complementary with other sensors. Radar-based human activity recognition (HAR) has the advantages of privacy preservation and noise robustness. Therefore, it has a tremendous potential for IoT applications. Existing methods are dedicate to finding best radar representations for HAR. However, research on the elimination of nonactivity information has not receive sufficient attention. In response to this question, the crucial region search (CRS) and feature discrimination (CRSFD) network has been proposed. It aims to automatically separate features and explicitly establish feature distribution, to accomplish nonactivity feature elimination. The CRSFD consists of the CRS module and the activity-associated feature discrimination (AAFD) module. The CRS module, with edge protection and importance assessment capabilities, is more suitable for irregular changes in activity features. The AAFD module is customized to automatically and explicitly analyze the distributions of activity and nonactivity features. Eventually, activity features are maintained for HAR. Experimental results on a human activity dataset and a human gesture dataset show that the proposed method has superior performance and is robust to noise.
7 November 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Deep learning methods have achieved huge success in synthetic aperture radar (SAR) target recognition, yet insufficient data limits their performances. If the key prior knowledge that describes the target is incorporated into the network, this problem can be effectively mitigated. Fortunately, electromagnetic scattering feature (ESF) elegantly and accurately models SAR targets, which describe the intrinsic properties of targets in SAR images. Therefore, how to perceive and integrate ESFs emerges as an essential issue. To address this problem, the adaptive scattering feature awareness and fusion network is proposed to obtain the more holistic target description. Specifically, the scattering structure intelligent awareness module is designed to automatically distinguish and screen the critical scattering information from the SAR image. Subsequently, in the feature fusion phase, the adaptive feature fusion module fuses the scattering domain ESFs and image domain deep features deeply and hierarchically by narrowing the distribution gap between them. In addition, a novel azimuth angle-guided classifier is employed to capture the association information between azimuth angle and target features. The physical laws inherent in ESF and azimuth angle guide the network training period, which improves feature discrimination and mitigates the pressure of data inadequacy. Experimental results on the moving and stationary target acquisition and recognition dataset and full aspect stationary targets-vehicle dataset demonstrate that adaptive scattering feature awareness and fusion network achieves state-of-the-art performance in SAR target recognition tasks.
Deep learning methods have been widely used in the field of synthetic aperture radar (SAR) target recognition. However, given the difficulty in obtaining high-quality SAR images, existing models tend to focus on non-target regions, leading to uncontrollable overfitting phenomena. To cope with such inherent problem, a novel global discriminative information search and focus (GDI-SF) network is proposed. The proposed framework obtains the holistic and pure description of the target without increasing the extra model parameters and annotations. Specifically, to capture the global description of the target, we employ higher-order self-correlation to enhance the interaction among features and elegantly aggregate global target-related information during the training period. In the view of the special imaging mechanism and scattering characteristics of SAR images, the images contain complex interference information, which will be coupled with the target features during the feature global interaction and fail to be separated easily. Thus, we constrain the input data to converge to the target region and eliminates the influence of target-irrelevant information from the source input. Under the above losses constraint, purer global discriminative target features are captured to yield more robust and superior recognition results elegantly. Finally, we conduct experiments on FAST-Vehicle dataset and SAR-ACD dataset, to verify the superior performance of the proposed method.
Synthetic aperture radar (SAR) target recognition tasks face the dilemma of limited training samples. The fusion of target scattering features improves the ability of the network to perceive discriminative information and reduces the dependence on training samples. However, existing methods are inadequate in utilizing and fusing target scattering information, which limits the development of target recognition. To address the above issues, the multi-level structured scattering feature fusion network is proposed. Firstly, relying on the visual geometric structure of the target, the correlation between local scattering points is established to construct a more realistic target scattering structure. On this basis, the scattering association pyramid network is proposed to mine the multi-level structured scattering information of the target to achieve the full representation of the target scattering information. Subsequently, the discriminative information in the features is measured by the information entropy theory, and the results of the measurements are employed as weighting factors to achieve feature fusion. Additionally, the cosine space classifier is proposed to enhance the discriminative capability of features and the correlation with azimuth information. The effectiveness and superiority of the proposed method are verified on two publicly available SAR image target recognition datasets.