Development of an Efficient Algorithm for Quick Detection of Environmental Oil Spills

Donatus U. Onyishi and Godswill Ofualagba
Keywords: Crude oil spill and detection, sky and vegetation segmentation, extraction, entropy filtration.
Tropical Journal of Science and Technology 2020 1(1), 93-104. Published: June 29, 2020


Oil spill has been a global problem in both developed and developing oil producing countries, as large spills of oil may have substantial environmental impacts. Several methods have been employed for oil spill detection. These ranges from real-time remote surveillance by flying aircrafts with surveillance teams, to employment of various types of sensors. The sensors employed include visible sensors, infrared sensors, ultraviolet sensors, radar sensors, and laser fluorosensors. This paper presents the development an algorithm for detecting oil spill. The oil spill algorithm has the ability to positively detect and characterize crude oil spills in the images acquired with video cameras. All crude oil spill images used in the development and testing of the algorithm were obtained from Crude Oil Spill Imaging Database, which is publicly available to researchers and image analysts at A total of 42 images were selected and used in the testing of the spill detection algorithm. To effectively detect an oil spill, the algorithm goes through four major steps of sky and vegetation segmentation, homogeneity extraction, entropy filtration, and standard deviation thresholding. The algorithm successfully detected the crude oil spill in 36 of the 42 images picked randomly from the Imaging database, which contains 104 images. This translated to a sensitivity of 85.7%.