Reference spectral signature selection using density-based cluster for automatic oil spill detection in hyperspectral...

Area of Science:

• Remote Sensing
• Spectroscopy
• Environmental Monitoring

Background:

• Automatic oil spill detection from hyperspectral images is crucial for environmental protection.
• Accurate reference spectral signature selection is a key challenge in this process.
• Existing methods may lack robustness in identifying specific oil signatures.

Purpose of the Study:

• To develop an automated approach for selecting reference spectral signatures of oil spills.
• To enhance the accuracy and reliability of oil spill detection using hyperspectral imaging.
• To address the fundamental challenge of spectral signature identification in complex environments.

Main Methods:

• A novel approach utilizing density-based clustering for reference spectral signature selection.
• Estimation of background parameters using seawater reflectance in infrared bands.
• Implementation of the conventional adaptive cosine estimator (ACE) algorithm for target detection.

Main Results:

• The proposed method successfully automates the selection of oil spill spectral signatures.
• High detection performance was achieved in experiments using real-world hyperspectral data.
• The approach demonstrated effectiveness in identifying oil spills from the Horizon Deep water incident.

Conclusions:

• Density-based clustering offers a robust solution for reference spectral signature selection.
• The developed framework significantly improves automated oil spill detection capabilities.
• This method provides a valuable tool for environmental monitoring and rapid response to oil spills.