An inproved binary encoding algorithm for classification of hyperspectral images

Binary encoding is a standard technique in classifying hyperspectral images. In this paper, an improved binary encoding (IBE) approach is proposed to integrate spectra, texture, shape and height (if it is possible) information of hyperspectral image data into a binary encoding algorithm for automatically deriving class cover information. First connected regions are extracted from the hyperspectral data by applying a segmentation algorithm. The mean spectrum per region is considered representative for the region. Predefined parameters were used to describing the texture and shape of the region. Together with the spectral information these parameters and the corresponding height values from the height source are converted into a binary code. This code is then matched to that of a training data set for classification.