Radiometric normalization for multitemporal and multispectral high resolution satellite images using ordinal conversion

Radiometric normalization of multi-temporal satellite image is very important for change detection or image mosaic. In this paper, a simple and effective radiometric normalization approach is proposed for processing high resolution multispectral satellite images such as IKONOS or QuickBird. This approach involves three main steps: (1) extracting pixel values from image pairs, (2) sorting original pixel values of image pairs in ascending order, and (3) assigning each pixel a new value based on its reflectance value relative to all other pixels. The effectiveness of the proposed approach is confirmed with experimental results compared to other normalization methods.