Oil Spill Detection in Gulf of Thailand using Normalized Difference Index with Local Mean Reflectance

The Gulf of Thailand is a region rich in activities. Oil drilling and transportation are an important one that plays a crucial role in the ASIAN economy. However, oil spills pose serious environmental risks and can impact human health both directly and indirectly. To improve disaster planning and management for oil spills, effective detection, monitoring, and mapping of spill areas are essential. However, detecting oil spills is a challenging task influenced by various factors, including weather conditions, observation angles, oil types, water depth and etc. Moreover, the interaction between spilled oil and different wavelengths of the electromagnetic spectrum varies, with some wavelengths performing effectively in specific conditions while failing in others. To address these challenges, this research integrates data across all wavelength ranges to enhance the accuracy and reliability of oil spill detection. This paper proposes an oil spill detection method based on the Normalized Difference Ratio. The method employs the local mean reflectance of each spectral band to calculate an individual Normalized Difference Index (NDI) relative to its surroundings. The NDIs from all bands are then combined to capture comprehensive oil spill information across the spectral range. The proposed method was evaluated using data from two sites in the Gulf of Thailand and compared with existing techniques. Both visual and quantitative results demonstrate that the method outperforms traditional approaches in detecting and mapping oil spills. This advancement offers a valuable tool for improving oil spill response and management, helping to mitigate environmental and economic damages in sensitive marine ecosystems.