Research on the categorical scale in remote sensing classification

This paper studies on the categorical scale of land cover class in remotely sensed imagery. On one hand, the categorical scale is referred to as the hierarchy of class. Different levels of classification represent various categorical scales. On the other hand, the categorical scale is used in specific class extraction, where the interest class may be extracted at different categorical scales. We give three measures for the categorical scale to quantify the classes in scale information. An experiment of categorical scale measure is performed in class extraction from remotely sensed image.