Monitoring oil spill thickness and weathering using UAV-borne hyperspectral sensing

Remote sensing plays a crucial role in the detection and monitoring of oil spills at sea, aiding in emergency response planning and monitoring routines. Despite significant progress in this field, estimating oil thickness remains challenging, impacting response efficacy and liability determination. Traditional methods like the Bonn Agreement Oil Appearance Code (BAOAC) are commonly used operationally. However, interpretation based only on visual identification suffers from subjectivity and limitations in the code designation for oil films. Hyperspectral sensing offers a promising solution for this matter. Previous works have demonstrated a relationship between oil spectral features and oil thickness. However, most of these studies rely on mathematical models validated with aerial or orbital data from old spills, hindering real-time applicability. This study proposes a novel approach based on unmanned aerial vehicles (UAVs) equipped with a hyperspectral camera for real-time data acquisition during offshore spills. The goal was to improve the BAOAC approach by integrating VNIR-SWIR spectral data with visual interpretation to classify oil spills. The methodology evaluates surface area, oil thickness, and weathering stage estimation through a realistic 11-week oil spill experiment. Results demonstrate that the high resolution of data acquired with UAVs can facilitate close-range monitoring of oil spills, especially near oil rigs and shorelines, providing more reliable results when combined with the BAOAC. The methodology can be upscaled by mounting the hyperspectral camera on aircraft or helicopters, providing promising applications for oil spill monitoring and offering practical insights for the short- and long-term management efforts carried out in the petroleum industry.