Image Processing in Polarimetric SAR Images Using a Hybrid Entropy Decomposition and Maximum Likelihood (EDML)

This paper presents a hybrid entropy decomposition and maximum likelihood (EDML) for synthetic aperture radar (SAR) image analysis and classification with the application of land use management. Entropy decomposition is an effective technique to obtain valuable decomposed parameters for image interpretation with analysis of the underlying scattering mechanisms. However, the main disadvantage of entropy decomposition is that the decision boundaries of the analysis plane are arbitrary. To overcome this problem, maximum likelihood technique is taken into consideration to help to determine the decision boundaries based upon Gaussian probability model. Hence, the hybrid EDML is developed to provide alternative way to improve the classification accuracy. The objective of this paper is to assess the use of polarimetric data for the image analysis and classification of land use management. It is illustrated using a well-known polarimetric AIRSAR data of San Francisco. In this paper, EDML technique is shown to have a superior result compared to other techniques.