Abstract

Images are perspective when they are captured. For applications like image based rendering and metrology from single view, the image will be perspective in nature. Generally, it requires parallel view image for photo realistic results. To get the photo realistic output, the perspective distortion of the image has to be rectified. We propose a method based on perspective transformation and plane homography to rectify the perspective distortion in an image to its actual scale with known camera parameters. Also, we study the amount of perspective distortion based on Hough Transform and the distance of vanishing points from the image boundary. The distance between the parallel lines in the perspective image decreases towards the vanishing point and in rectified image it remains constant. The results produced out of the proposed methods are photo realistic and more suitable for applications like image based rendering.

Keywords: Rectification, perspective image, Vanishing point, plane homography, Hough transform, rendering application, camera calibration.
(P1150721004, 1.8 MB)

	Mohan.S, got his B.E in computer Science and Engineering in 1998 from Bharathidasan university, Tamilnadu, India and M.Tech in computer cognition technology in 2002 from University of Mysore, India. Presently working towards his PhD on Image based modelling and rendering in P.E.T research centre, PESCE, Mandya, Karnataka India affiliated to University of Mysore. He is working as Technical Evangelist in Education & Research Department, Infosys Technologies ltd, Mysore, India.
	Avinash N, got his B.E in 2001 and M.Tech computer cognition technology in 2003 from University of mysore, India. Currently he is working towards his PhD on Vision based metrology in P.E.T research centre, PESCE, Mandya, Karnataka India affiliated to University of Mysore.
	Prof.Murali.S, working as professor and head of Information science and Engg department in PESCE, Mandya, India. He got his PhD from University of Mysore, in 2002. His research interests include Computer Vision,  Document Imaging and pattern recognition.

@ARTICLE{P1150721004,

AUTHOR = {Mohan  S and Avinash  N. and Murali.S},

TITLE = {Rectification  of  Perspective  distortion  using  camera  parameters - A  Perspective  Geometry  Based  Approach},

JOURNAL ={ICGST International Journal on Graphics, Vision and Image Processing, GVIP},

YEAR = {2008},

VOLUME = {08},

ISSUE ={I},

PAGES={1--7}

}

(P1150721004, 1.8 MB)