Graph neural network (GNN) has recently gained increasing attention in the hyperspectral image (HSI) classification. Compared with convolutional neural network (CNN), GNN can effectively relieve the scarcity of labeled data. In our method, we first perform feature learning on large-scale irregular regions through GNN and then extract local spatial–spectral features at the pixel level. Besides, we incorporate edge convolution (EdgeConv) into GNN to adaptively capture the interrelationship of the representative descriptors and fully exploit the discriminative features on graph. Experiments on several HSI datasets show that our method can achieve better classification performance compared with the state-of-the-art HSI classification methods.

This study proposes an asymmetric mapping method to design sparse planar arrays with multiple constraints including the aperture, number of elements and minimum spacing between adjacent elements. Benefiting from two different mapping functions, which are established in this paper, this approach allows one to overcome the limitations of the existing matrix mapping method in terms of flexibility and performances. The numerical validation points out that the proposed method outperforms the matrix mapping method and modified real genetic algorithm in the design of array arrangements.

Semisupervised band selection (BS) technique plays an important role in processing hyperspectral images (HSIs) because of its superiority of using the limited labeled data and plentiful unlabeled data to select the discriminative and informative feature, which copes with the problem of high-dimensional and scare labeled samples of HSIs. Among semisupervised BS models, graph-based methods are superior to others in many situations and have received more and more attention. However, traditional graph-based models construct a similarity matrix and select valuable bands independently. The similarity matrix remains constant, which will damage the local manifold structure and lead to a suboptimal result. To solve this problem, we propose a semisupervised band selection with an optimal graph (BSOG) approach, which performs BS and local structure learning simultaneously. Instead of fixing the input similarity matrix, the similarity matrix is updated constantly to learn a better local structure. Besides, the learned similarity matrix is adaptive. Then, the optimal band subset can be selected by analyzing the obtained projection matrix $\boldsymbol W$ . An efficient and simple optimization algorithm is proposed to solve this model. Experiments on four real HSIs data validate the effectiveness of the proposed model.

Graph neural networks (GNNs) have been widely applied for hyperspectral image (HSI) classification, due to their impressive representation ability. It is well-known that typical GNNs and their variants work under the assumption of homophily, while most existing GNN-based HSI classification methods neglect the heterophily that is widely present in the constructed graph structure. To deal with this problem, a homophily-guided Bi-Kernel Graph Neural Network (BKGNN) is developed for HSI classification. In the proposed BKGNN, we estimate the homophily between node pairs according to a learnable homophily degree matrix, which is then applied to change the propagation mechanism by adaptively selecting two different kernels to capture homophily and heterophily information. Meanwhile, the learning process of the homophily degree matrix and the bi-kernel feature propagation process are trained jointly to enhance each other in an end-to-end fashion. Extensive experiments on three public data sets demonstrate the effectiveness of the proposed method.

In this paper, we investigate the physical layer security of cooperative two-way relay transmission systems using the amplify-and-forward (AF) protocol in the presence of an eavesdropper. A joint relay selection (RS) and power allocation (PA) scheme is proposed to protect the source-destination transmission against the eavesdropper. However, due to the high computational complexity, it is difficult to obtain the optimal solution for the system secrecy rate. Fortunately, an approximate optimal solution by using the particle swarm optimization (PSO) algorithm is derived. In the simulations, we use random relay selection with optimal power allocation (RRS-OPA) and equal power allocation with optimal relay selection (EPA-ORS) as benchmark schemes to verify the effectiveness of the proposed method. The simulation results show that the proposed method outperforms both RRS-OPA and EPA-ORS and significantly improves the system performance with low complexity.

This paper details a method to improve accuracy of the land vehicle attitude estimation. It employs the vehicle model constraints to enhance the integration of GPS and a low-cost MEMS inertial measurement unit (MIMU) (GPS/MIMU). To improve the yaw angle estimation, we propose a lateral velocity constraint (LVC) aided method based on the observability analysis for the integrated fusion system. The theoretical analysis indicates that the yaw angle will be directly observed if LVC is augmented into the GPS/MIMU fusion algorithm. Furthermore, for LVC/MIMU system without GPS, the roll angle is always well estimated regardless of whether the vehicle is moving or stationary, and, surprisingly, the pitch angle can also be relatively accurately estimated if the vehicle is moving. Thus, when GPS is unavailable, the LVC/MIMU fusion is proposed by replacing the GPS measurement with the virtual measurement LVC. It is an encouraging improvement since the accumulating errors induced by gyro bias can be suppressed using MIMU alone. Moreover, to overcome the shortage that the pitch angle is not observed if the vehicle is stationary, we propose to fuse the gyros with gravity-based angles by Kalman filtering if the vehicle is detected motionless based on a switching rule. The experimental results compare favorably to theoretical analysis, showing that the proposed method can effectively improve the accuracy for land vehicle attitude estimation.