Feature Extraction of Offshore Oil Slick from Hyperspectral Data Based on EMD and Neural Network
基于EMD与神经网络的油膜高光谱数据特征提取

Based on the reflectance spectrum of oil slick and background seawater,an EMD based approach to feature extraction from hyperspectral data is presented in the work,and a BP neural network is constructed to classify the oil slick and seawater samples.Firstly,the EMD is adopted to decompose the hyperspectral data from 425 nm to 2 390 nm into a series of Intrinsic Mode Functions ( IMFs) ,and the energy of each IMFs is computed,then based on the computed energy,some IMFs energy are selected as inputed parameter of neural networks to classify the oil slick and seawater.The results show that the EMD based feature extraction approach can identify oil slick accurately and effectively.