期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2008
卷号：XXXVII Part B2
页码：385-390
出版社：Copernicus Publications
摘要：In this paper, a new approach is presented to select the optimal scale for image classification obtaining desirable overall classification accuracy. The objective of the new approach is to choose optimal scale with minimal average entropy of separability. The concept of information entropy is introduced to quantify the uncertainty in image classification. Moreover, the spatial distribution is also taken into account for the choice of optimal scale. Since local variance, variogram and transformed divergence are introduced to select optimal scale in many cases, we design a comparative experiment to compare our approach with the existing ones. By using an image sequence with different spatial resolutions, the experiment show that the new approach is more meaningful than traditional ones, thus benefiting the choice of optimal scale in image classification and many other applications
