The Potential of Multi- and HyperSpectral Air- and Spaceborne Sensors to Detect Crude Oil Hydrocarbon in Soils Long after a Contamination Event

Ran Pelta,Eyal Ben-Dor

doi:10.3390/app9235151

Abstract

Crude oil contamination is hazardous to health, negatively impacts vital life sources, and causes land and ecological degradation. The basic premise of the prevalent spectroscopic analyses for detecting such contamination is that crude oil spectral features are observable in the spectrum. Such analyses, however, have failed to address instances where the expected spectral features are not visible in the spectrum. Hence, a more refined method was recently published, which accounts for such cases. This method was successfully applied to a hyperspectral image over an arid area long after a contamination event. This study aimed to determine whether that same method could be successfully applied using a variety of other operational and future instruments, both air- and spaceborne, with different spatial and spectral characteristics. To that end, a series of simulation experiments was performed, including various spectral and spatial resolutions. Quantitative and qualitative evaluations of the classification are reported. The results indicate that the hyperspectral information can be reduced to one-third of its original size, while maintaining high accuracy and a quality classification map. A ground sampling distance of 7.5 m seems to be the boundary of an acceptable classification outcome. The overall conclusion of this study was that the method is robust enough to perform under various spectral and spatial configurations. Therefore, it could be a promising tool to be integrated into environmental protection and resource management programs.

Highlights

Anthropogenic crude oil pollution is a global problem, endangering many crucial natural resources, such as drinking water and arable land, as well as causing environmental damage in diverse ecological systems
The detection of crude oil pollution on terrestrial surfaces using hyperspectral remote sensing relies primarily on the chemical structure of the organic compounds found in crude oil, termed petroleum hydrocarbons (PHCs)
Because of its chemical structure (C–H bonds), PHCs give rise to spectral manifestations in the shortwave infrared (SWIR; 1000–2500 nm) portion of the electromagnetic spectrum, at around 1200, 1700, and 2300 nm [1]. These PHC spectral features, can be detected mainly by utilizing hyperspectral sensors, which record the amount of reflected energy in tens or hundreds of continuous narrow spectral bands on a spatial domain, allowing for the identification of the PHC-SF

Summary

Introduction

Anthropogenic crude oil pollution is a global problem, endangering many crucial natural resources, such as drinking water and arable land, as well as causing environmental damage in diverse ecological systems. The detection of crude oil pollution on terrestrial surfaces using hyperspectral remote sensing relies primarily on the chemical structure of the organic compounds found in crude oil, termed petroleum hydrocarbons (PHCs). Because of its chemical structure (C–H bonds), PHCs give rise to spectral manifestations in the shortwave infrared (SWIR; 1000–2500 nm) portion of the electromagnetic spectrum, at around 1200, 1700, and 2300 nm [1]. These PHC spectral features (hereafter termed PHC-SF), can be detected mainly by utilizing hyperspectral sensors, which record the amount of reflected energy in tens or hundreds of continuous narrow spectral bands on a spatial domain, allowing for the identification of the PHC-SF. Pabón et al [6] demonstrated that the Advanced Spaceborne

Objectives

Methods

Discussion

Conclusion