Volume IV-5
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., IV-5, 293-301, 2018
https://doi.org/10.5194/isprs-annals-IV-5-293-2018
© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., IV-5, 293-301, 2018
https://doi.org/10.5194/isprs-annals-IV-5-293-2018
© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

  15 Nov 2018

15 Nov 2018

OIL SPILLS DETECTION BY MEANS OF UAS AND LOW-COST AIRBORNE THERMAL SENSORS

A. Al-Shammari, E. Levin, and R. Shults A. Al-Shammari et al.
  • Michigan Technological University, School of Technology, Townsend Drive, 1400 Houghton, 49931, USA

Keywords: oil spills, monitoring, thermal band, UAS, image classification

Abstract. This paper provides an overview of oil spill scenarios and the remote sensing methods used for detection and mapping the spills. It also discusses the different kinds of thermal sensors used in oil spills detection. As UAS is becoming an important player in the oil and gas industry for the low operating costs involved, this research involved working with a cheap thermal airborne sensor mounted on DJI Phantom 4 system. Data were collected in two scenarios, first scenario is collecting data in Michigan’s Upper Peninsula at a petroleum company location and the second scenario was an indoor experiment simulating an offshore spill. The aim of this research is to inspect the capability of Lepton LWIR inexpensive sensor to detect the areas contaminated with oil. Data processing to create classification maps involved using ArcGIS 10.5.1, ERDAS Imagine 2015 and ENVI 5.3. Depending accuracy assessment (confusion matrices) for the classified images and comparing classified images with ground truth, results shows the Lepton thermal sensor worked well in differentiating oil from water and was not a good option when there are many objects in the area of interest. Future research recommendations and conclusions are presented.