Enhanced IHS Pan-sharpening using K-Means Segmentation Guided Adaptive Intensity Histogram Matching and CLAHE Enhancement

Pan-sharpening is the method of combining a panchromatic image of higher spatial resolution with a multispectral image of lower spatial resolution. The main aim of pan-sharpening method is that the spectral characteristics of multispectral data should be preserved with maximum spatial enhancement. It is observed that Component Substitution (CS) based pan-sharpening techniques such as Brovey and Intensity, Hue and Saturation (IHS) transform inject the spatial details but suffers from large spectral distortion. To overcome this problem, we introduced a novel method that enhance IHS pan sharpening and generate a representative image that has optimal spatial and spectral resolution. The K-Means segmentation of the multispectral image describe the type of local feature patch and guide the filtering method to generate adaptive intensity component. The panchromatic band is histogram matched with adaptive intensity and inverse IHS transform to generate K-means guided fused image. In addition, the fused image is enhanced using Contrast Limited Adaptive Histogram Equalization (CLAHE) for better feature perceptibility. The proposed method is evaluated with IKONOS panchromatic and multispectral concurrent imagery covering diverse spectral targets. The visual quality assessment indicates that enhanced IHS clearly demark the urban structures. The quantitative evaluation with different image quality metrics shows that our proposed methodology performs better than state-of-the-art CS pan-sharpening techniques.