Oil Spill Detection on Landsat-8/9 Images Based on Deep Learning Methods

Abstract

Oil pollution of seas and oceans poses a danger to human health and has a big impact on the marine environment. Oil enters the water from various sources which are of natural (47 %) and anthropogenic origin (53 %). Human-caused oil pollution is mainly linked to the progressive increase in oil production, consumption and transportation, as well as the general increase in the transportation of goods by sea. The most common cases are accidents in maritime transportation, on oil platforms or deliberate discharges of oil from ships where large amounts of oil can be released into the water during a short time. A timely and accurate oil detection can help to prevent pollution spread and support clean-up operations to minimize the negative impacts on the environment as well as to identify the polluter. Remote sensing has been proven to be effective for monitoring of large areas. This paper presents two different approaches for automated oil spill detection from multispectral satellite images based on a deep neural network (DNN) and the convolutional neural network (CNN). The presented results are based on a very small number of satellite images acquired with the optical satellites Landsat-8 and Landsat-9.