Vanishing point and vanishing line estimation with line clustering

Abstract

Introduction The vanishingp oint is defined as the convergencep oint of lines in an image page that is p: duced by the pe: jection ofp:FE6W: lines in realspl:5 The vector from the center of a camera lens to a vanishingp oint refers to the vector ofp5LL6W: lines. Therefore, vanishingp oint detection is used to extract three-dimensional scenes. The pe:6S5 of detection is researched as an intersectionp oint detectionpectio in the image page that is p: duced by agroup ofp:56LEF lines in realspl:S However, due to quantization and line detection error, these lines do not always intersect. Therefore, several methods have beenpn:5( ted to address these pese lems. These methods can be divided into two categories: the votingapg:5S h using the Hough Transform or a Gaussianspsia and the statisticalapatis h. Kenderpnd osed a method of the former typ e that Manuscript received May 12, 1999. Manuscript revised December 3, 1999. + The authors are with the Graduate Sch