Extraction of Oil Spill Information Using Decision Tree Based Minimum Noise Fraction Transform

Abstract

In order to reduce the number of bands for processing hyperspectral remote sensing data and to improve the processing efficiency, this article proposed a decision tree classification method based on minimum noise fraction (MNF) transform. MNF transform was used to reduce data redundancy, and the image noise was separated. By analyzing the MNF eigenvalues of the ground objects, the classification decision tree was established, and the information such as the relative thickness of the oil film was extracted. The results show that the method can ensure recognition accuracy, and achieve the efficient use of information of spectral dimension. Meanwhile, the data processing time is significantly reduced, which is very important during emergency response to oil spills.