Compared with natural scenes, aerial scenes are usually composed of numerous objects densely distributed within the aerial view, and thus more key local semantic features are needed to describe them. However, when existing CNNs are used for remote sensing image classification, they typically focus on the global semantic features of the image, and especially for deep models, shallow and intermediate features are easily lost. This paper proposes a channel-spatial attention mechanism based on a depthwise separable convolution (CSDS) network for aerial scene classification to solve these challenges. First, we construct a depthwise separable convolution (DS-Conv) and pyramid residual connection architecture.DS-Conv extracts features from each channel and merges them, effectively reducing the number of necessary calculations, and the pyramid residual connections connect the features from multiple layers and create associations. Then, the channel-spatial attention algorithm causes the model to obtain more effective features in the channel and spatial domains. Finally, an improved cross-entropy loss function is used to reduce the impact of similar categories on backpropagation. Comparative experiments on three public data sets show that the CSDS network can achieve results comparable to those of other state-of-the-art methods. In addition, visualization of feature extraction results by the Grad-CAM algorithm and ablation experiments for each module reflect the powerful feature learning and representation capabilities of the proposed CSDS network.
Accurate localization of nodes is one of the key issues of wireless sensor network (WSN). A localization algorithm using expected hop progress (LAEP) has been successfully applied in isotropic wireless sensor networks. However, range-free LAEP cannot be directly used for anisotropic WSNs because anisotropic problems limit the applicability of multi-hop localization. In order to solve the problem, an improved localization algorithm is proposed to reduce the localization error. In this paper, we adapt the expected hop progress to anisotropic WSNs by considering both hop count computation and anchor selection. Then, particle swarm optimization algorithm is introduced to improve the positioning accuracy. The experimental results demonstrate that our algorithm has better higher precision than do state-of-the-art algorithms. Even for isotropic WSNs, our algorithm always outperforms its counterparts.
The segmentation of synthetic aperture radar (SAR) water-land images is a very difficult task not only because of strong multiplicative noise but also due to the blurred boundary, irregular shape, and together with diminished contrast. In this paper, we propose a matrix factorization active contour model based on fused features for SAR image segmentation. First, to enhance the robustness, multiple features are utilized. For each pixel location, feature maps (matrix) are constructed by combining wavelet textual features, Gaussian (DoG) filter features, and Gabor filter features via local spectral histogram, which improves spatial pattern and express image structure. Second, the energy function is constructed based on region information and edge information of SAR image. Region information is obtained via matrix factorization theory on the feature matrix. Edge information is obtained by modified the ratio of exponentially weighted averages operator. Then, a convex energy function is proposed to avoid the local minima. A fast dual formulation is introduced for the evolution of the contour. Finally, synthetic and real SAR data are used for verification. The experimental results demonstrate the proposed algorithm is effective for water/land segmentation in SAR images. INDEX TERMS Synthetic aperture radar (SAR), remote sensing, image segmentation, active contour, edge detector, matrix decomposition techniques, texture feature.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.