Title :
A New Method for Ellipse Fitting in Satellite Images
Author :
Soh, Youngsung ; Bae, Jongwoo ; Kim, Daewon ; Kim, Heejun
Author_Institution :
Dept. of Inf. Eng., Myongji Univ., Yongin, South Korea
Abstract :
A new method for ellipse fitting in satellite image is proposed. First, line segments are extracted and classified into one of three categories: straight line, elliptic line, and pseudo-elliptic line segments. Then elliptic segments are fitted directly with non-iterative ellipse-specific method proposed by Fitzgibbon et al[12]. For pseudo-elliptic segments, we start fitting ellipse with large eccentricity, merge nearby pseudo-elliptic segments that share same monotonicity, validate fitted ellipse, and accept if validation succeeds. If validation fails, fitting process is repeated by lowering eccentricity until either validation succeeds or eccentricity hits zero. We tested the method on some satellite image and obtained good results.
Keywords :
feature extraction; image classification; least squares approximations; ellipse fitting; elliptic line segments; noniterative ellipse-specific method; pseudoelliptic line segment; satellite images; straight line segments; Automation; Buildings; Data mining; Equations; Image analysis; Image segmentation; Quantization; Roads; Satellites; Testing; ellipse fitting; line segment extraction;
Conference_Titel :
Intelligent Computation Technology and Automation, 2009. ICICTA '09. Second International Conference on
Conference_Location :
Changsha, Hunan
Print_ISBN :
978-0-7695-3804-4
DOI :
10.1109/ICICTA.2009.128