A Deep Learning Approach to Detecting Ships from High-Resolution Aerial Remote Sensing Images

Abstract

Huang, B.; He, B.; Wu, L., and Lin, Y., 2020. A deep learning approach to detecting ships from high-resolution aerial remote sensing images. In: Liu, X. and Zhao, L. (eds.), Today's Modern Coastal Society: Technical and Sociological Aspects of Coastal Research. Journal of Coastal Research, Special Issue No. 111, pp. 16–20. Coconut Creek (Florida), ISSN 0749-0208.Ship activity has become progressive frequent with maritime safety's evolution. The accurate and fast localization of marine ships is therefore growing increasingly important. Because high-resolution aerial remote sensing (HRARS) images have a high spatial resolution, so automatic ship recognition in HRARS images has gained extensive attention for its vast applications. However, identifying ships from interferences, such as features of clouds, waves, and some land architectures that are similar to ships, is difficult. In this paper, an efficient and robust algorithm to detect ship from HRARS images is proposed. Our ship-detection framework is designed based on a typical deep learning algorithm, which provides exact recognition and locations of ship targets in an automatic feature extraction method. Experimental results on a large quantity of high-resolution images verify that the ship objective's characteristics are represented with effect and the detection performance is significantly promoted.