Vector Ordering and Multispectral Morphological Image Processing

摘要: This chapter illustrates the suitability of recent multivariate ordering approaches to morphological analysis colour and multispectral images working on their vector representation. On one hand, supervised renders machine learning notions image processing techniques, through a stage provide total in colour/multispectral space. other anomaly-based ordering, automatically detects spectral diversity over majority background, allowing an adaptive salient parts image. These two paradigms allow definition operators for images, from algebraic dilation erosion more advanced techniques as simplification, decomposition segmentation. A number applications are reviewed implementation issues discussed detail.