Scene Classification With Recurrent Attention of VHR Remote Sensing Images

Scene classification of remote sensing images has drawn great attention because of its wide applications. In this paper, with the guidance of the human visual system (HVS), we explore the attention mechanism and propose a novel end-to-end attention recurrent convolutional network (ARCNet) for scene classification. It can learn to focus selectively on some key regions or locations and just process them at high-level features, thereby discarding the noncritical information and promoting the classification performance. The contributions of this paper are threefold. First, we design a novel recurrent attention structure to squeeze high-level semantic and spatial features into several simplex vectors for the reduction of learning parameters. Second, an end-to-end network named ARCNet is proposed to adaptively select a series of attention regions and then to generate powerful predictions by learning to process them sequentially. Third, we construct a new data set named OPTIMAL-31, which contains more categories than popular data sets and gives researchers an extra platform to validate their algorithms. The experimental results demonstrate that our model makes great promotion in comparison with the state-of-the-art approaches.

I. Introduction

With the rapid development of remote sensing instruments over recent years [1], [2], very high-resolution (VHR) remote sensing images are becoming increasingly available and bringing us the opportunity to try more research in military and civilian applications, such as natural disaster detection [3], [4], land-cover/land-use classification [5], [6] geographic space object detection [7], [8], geographic image retrieval [9], [10], urban planning, and environment monitoring. As we all know, VHR remote sensing images recognition based on the knowledge of domain experts has a high labor cost. Therefore, intelligent scene classification of remote sensing images [11]–[14], which categorizes scene images into different classes based on its semantic information, has drawn great attention in remote sensing field. Nevertheless, because of various classes of scenes and complex spatial information of VHR remote sensing images, how to effectively describe and classify the scenes is a pivotal and challenging task.