Please use this identifier to cite or link to this item: https://www.um.edu.mt/library/oar/handle/123456789/11401
Title: An automatic keystone correction system for projectors
Authors: Gauci, Jean
Keywords: Image processing
Geometry, Projective
Projectors
Issue Date: 2015
Abstract: As researchers find new ways of making electronic devices smaller, consumer devices have seen a drastic decrease in size over the years. Today handheld projectors and mobile devices with embedded projectors are a reality. Moreover, immersive projections and projection mosaics are widely implemented in various scenarios. This, coupled with the traditional use of projectors in presentation rooms makes projectors one of the most researched consumer devices on the market. Keystone distortion is an inherent problem in all kinds of projectors. Keystone distortion occurs when the projection axis of the projector is not parallel to the projection screen and thus the projected content appears as an arbitrary quadrangle rather than a rectangle. An automatic keystone distortion correction algorithm should correct this distortion without any user input and should correct the distortion on any arbitrary blank screen. In other words, the algorithm shouldn’t rely on any screen markings. Furthermore the algorithm should carry out the corrections continuously since the projector might be continuously moving, as is the case in mobile projectors. In this project an automatic keystone distortion correction algorithm is developed by using a camera attached to a projector and a computer for computation of the image warping. The use of a camera is not seen as a limitation since devices released in the market increasingly incorporate both. The algorithm first detects the projection region in the camera image and works out the largest, non-distorted, aligned rectangle inside which to fit the new projection region. The projector image is then pre-warped in position. The procedure is repeated for continuous correction on the fly. In order to do so, the system relies on the detection of a projection screen and so the user must project to such a screen and not any arbitrary flat surface. Testing results have shown that the recovered projection region is really a world aligned, correct aspect ratio rectangle and the system performs the calculations and correction in real time.
Description: B.SC.IT(HONS)
URI: https://www.um.edu.mt/library/oar//handle/123456789/11401
Appears in Collections:Dissertations - FacICT - 2015
Dissertations - FacICTCCE - 2015

Files in This Item:
File Description SizeFormat 
15BSCIT062.pdf
  Restricted Access
2.43 MBAdobe PDFView/Open Request a copy


Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.