作者: Du-Ming Tsai , H.-T. Hou , H.-J. Su
DOI: 10.1016/S0167-8655(98)00130-5
关键词: Corner detection 、 Covariance 、 Mathematics 、 Covariance matrix 、 Topology 、 Boundary (topology) 、 Curvature 、 Mathematical analysis 、 Eigenvalues and eigenvectors 、 Measure (mathematics) 、 Estimation of covariance matrices
摘要: Abstract In this paper we present a new measure for corner detection based on the eigenvalues of covariance matrix boundary points over small region support. It avoids false alarms superfluous corners circular arcs. Experimental results have shown that proposed methods using curvature measures good and localization curved objects in different rotations with varying scale changes.