Advancing Hyperspectral Remote Sensing for Oil Contamination Detection and Characterisation in the Niger Delta

Hyperspectral remote sensing offers fundamentally superior capabilities for oil contamination detection and characterisation compared to conventional multispectral sensors, yet its operational application in the Niger Delta remains limited by data availability and methodological gaps. This study advances hyperspectral remote sensing methods for oil contamination detection and characterisation using airborne AVIRIS-NG imagery and laboratory hyperspectral measurements of contaminated soil and vegetation samples from the Niger Delta. Spectral libraries of fresh crude oil, weathered crude oil, emulsified oil, and oil-contaminated soil at varying concentrations are developed. Spectral mixture analysis and machine learning classification algorithms are applied to hyperspectral imagery to map contamination types and estimate contamination concentration levels. Novel spectral indices sensitive to crude oil absorption features at 1,210 nanometres and 1,730 nanometres are derived and evaluated. The methodology is validated against 84 geo-referenced field samples with measured total petroleum hydrocarbon concentrations, achieving a detection threshold of 2.3 percent oil by volume. The study provides a methodological foundation for deploying hyperspectral monitoring in support of the Ogoniland remediation programme. Keywords: hyperspectral, oil contamination, Niger Delta, spectral library, contamination mapping