Improved 3D Reconstruction using Stereo Vision with Gaussian Filter and Polynomial Disparity Fitting

  • 劉 冠甫

Student thesis: Master's Thesis


In this thesis it will be explained how the stereo vision can be used to calculate the distance between an object and a camera By using a back projection method in 3D space two existing images are recovered by which to analyze the contour of a target object Stereo matching is the most important step in stereo vision as it will determine the accuracy and computational time of required to obtain experimental results This thesis discusses both accuracy and computational time In terms of an algorithm the process is changed from FNCC matching to SSD matching Then a disparity range limitations is used to reduce both computational time and the number of mismatched points To further improve the accuracy a Gaussian filter and polynomial disparity fitting are added An unsharp mask filter is used in the experiment to sharpen images the window size is altered and different filters are used to find the best way to improve the accuracy and speed of stereo matching In conclusion propose method reduced both the number of mismatching points and computational time
Date of Award2016 Sep 5
Original languageEnglish
SupervisorTa-Chung Wang (Supervisor)

Cite this