Improved GIHSA for image fusion based on parameter optimization
Abstract:The most widely used image fusion method is based on the intensity-hue-saturation (IHS) transform, and many derivations have been developed from the basic IHS image fusion method. The recently developed generalized IHS adaptive (GIHSA) approach, which includes a parameter optimization procedure for the IHS transform based on a linear regression between panchromatic (Pan) and multispectral (MS) data, is among the best of the image fusion methods based on the IHS transform. However, GIHSA is only a semi-adaptive approach because only half of the parameters for fusion are adaptively determined in it. Therefore, we propose a fully adaptive image fusion approach under the general scheme, as an improvement for GIHSA. In the proposed approach, all parameters for fusion are adaptively determined following a two-step optimization procedure. The proposed approach is applied to Enhanced Thematic Mapper Plus (ETM+) images, and the improvements are verified by fusion results.
Document Type: Research Article
Affiliations: School of Electronic and Information Engineering, Dalian University of Technology, Ganjingzi Qu, Dalian, China
Publication date: March 1, 2010