Segmentation of High-resolution Multi-spectral Image of Urban Areas Based on Extended Morphological Profiles

High-resolution multi-spectral remote sensing image of urban areas provides both structural and spectral information about urban scenes. In segmentation of such complex image scenes, very thin, enveloped or nested regions may have to be retained. Standard morphological segmentation approaches which are based on edge-detection do, not perform well for such scenes. In this study, segmentation of such images based on extended morphological profiles is proposed. First, fundamental morphological vector operations (erosion and dilation) are defined by extension, taking into account the spatial and spectral information in simultaneous fashion. Theoretical definitions of extended morphological operations are used in the formal definition of the concept of extended morphological profiles, which is constructed based on the repeated use of openings and closings by reconstruction with a structuring element (SE) of increasing size. Then, the morphological multi-scale characteristic of the image at each pixel is defined as the SE size with the greatest associated value in the corresponding derivative of the extended morphological profiles. The multi-scale segmentation derived from the morphological multi-scale characteristic could not be the final segmentation result because of over- or under- segmentation in local parts of the image. Therefore, appropriate post-processing is used to process the previous multi-scale segmentation to gain more accurate segmentation result. The proposed approach is applied to high-resolution multi-spectral QUICKBIRD imagery of urban areas. The experiment result demonstrates good performance of this approach.