Abstract
Semantic segmentation is one of the significant tasks in understanding aerial images with high spatial resolution. Recently, Graph Neural Network (GNN) and attention mechanism have achieved excellent performance in semantic segmentation tasks in general images and been applied to aerial images. In this paper, we propose a novel Superpixel-based Attention Graph Neural Network (SAGNN) for semantic segmentation of high spatial resolution aerial images. A K-Nearest Neighbor (KNN) graph is constructed from our network for each image, where each node corresponds to a superpixel in the image and is associated with a hidden representation vector. On this basis, the initialization of the hidden representation vector is the appearance feature extracted by a unary Convolutional Neural Network (CNN) from the image. Moreover, relying on the attention mechanism and recursive functions, each node can update its hidden representation according to the current state and the incoming information from its neighbors. The final representation of each node is used to predict the semantic class of each superpixel. The attention mechanism enables graph nodes to differentially aggregate neighbor information, which can extract higher-quality features. Furthermore, the superpixels not only save computational resources, but also maintain object boundary to achieve more accurate predictions. The accuracy of our model on the Potsdam and Vaihingen public datasets exceeds all benchmark approaches, reaching 90.23% and 89.32%, respectively.
Highlights
With the rapid development in aerial photography technology in recent years, significant improvement has been achieved in spatial resolution of aerial images
Our method takes the complementary advantages of two neural networks based on superpixels to achieve satisfactory segmentation results
Ablation Study In Superpixel-based Attention Graph Neural Network (SAGNN), three important modules are used on the Graph Neural Network (GNN) body: superpixel module, attention module, and (CNN) feature extraction module, among which the superpixel module is used to reduce the resolution of the image and retain the boundary information of the object, the attention module is used to focus on similar neighbor information when clustering neighbors, and the Convolutional Neural Network (CNN) feature extraction module is used to extract feature vectors from the original image
Summary
With the rapid development in aerial photography technology in recent years, significant improvement has been achieved in spatial resolution of aerial images. Networks (GNNs) can process non-Euclidean structural data, effectively extract spatial features from topologies, and use global context information for inference learning [13,14]. To address the above challenges, a semantic segmentation method is proposed for aerial images based on superpixel-GNN with attention mechanism. A GNN-based framework has been proposed for semantic segmentation of aerial images. To solve the problem of irregular object edges in aerial images, superpixels are used as graph nodes to construct the graph structure, and GNN can directly learn its representation from the superpixel graph. Our method takes the complementary advantages of two neural networks (image features extracted by CNN and spatial relations provided by GNN) based on superpixels to achieve satisfactory segmentation results. The GNN model in our framework of semantic segmentation of aerial images is an improved version that has introduced the attention mechanism into each node.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
