Variance gain index for detection of boundary points in discrete one-dimensional data
Abstract:The authors propose a new algorithm for assisting the visual interpretation of remotely sensed images, in which a new statistical index, called the variance gain index, is used to extract boundary points from one-dimensional profile data. The index is described, and experimental examples are used to examine the feasibility of the method. This method is demonstrated to provide stable and accurate detection of boundary points in simulated one-dimensional data comprised of two sections.
Document Type: Research Article
Affiliations: 1: Department of Environmental Information, Faculty of Informatics, Tokyo University of Information Sciences, 1200-2, Yatoh-cho, Wakaba-ku, Chiba-city 265-8501, Japan; e-mail: firstname.lastname@example.org 2: Center for Spatial Information Science, University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo, 135-8508, Japan
Publication date: 2003-01-01