H. Hanaizumi et al., A BINARY DIVISION ALGORITHM FOR CLUSTERING REMOTELY-SENSED MULTISPECTRAL IMAGES, IEEE transactions on instrumentation and measurement, 44(3), 1995, pp. 759-763
In binary division clustering (BDC), image data are repeatedly divided
into two groups until the group consists of a single cluster, Canonic
al correlation analysis-is used for data compression and for noise red
uction, BDC achieved higher accuracy and higher efficiency than conven
tional ISODATA. BDC was successfully applied to change detection of re
motely sensed multitemporal multispectral images.