4.5

4.5 | 6 ratings Rate this file 108 Downloads (last 30 days) File Size: 2.31 MB File ID: #28522
image thumbnail

Stereo Matching

by Wim Abbeloos

 

20 Aug 2010 (Updated 27 Sep 2010)

A GUI demonstrates an efficient stereo matching algorithm.

| Watch this File

File Information
Description

 FAST MATLAB STEREO MATCHING ALGORITHM (SAD)

This function performs the computationally expensive step of matching two rectified and undistorted stereo images. The output is a dense disparity map. If camera parameters are known, this allows for three dimensional reconstruction.

Two graphical user interfaces demonstrate the algorithm.

Please note this function requires the Image Processing Toolbox!

[spdmap, dcost, pcost, wcost] = stereomatch(imgleft, imgright, windowsize, disparity, spacc)

The standard images included are from
[1] D. Scharstein and R. Szeliski. A taxonomy and evaluation of dense two-frame stereo correspondence algorithms. International Journal of Computer Vision, 47(1/2/3):7-42, April-June 2002.
[2] D. Scharstein and R. Szeliski. High-accuracy stereo depth maps using structured light. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2003), volume 1, pages 195-202, Madison, WI, June 2003.

Please do provide feedback.

Required Products Image Processing Toolbox
MATLAB release MATLAB 7.4 (R2007a)
Other requirements Image Processing Toolbox
Tags for This File  
Everyone's Tags
Tags I've Applied
Add New Tags Please login to tag files.
Comments and Ratings (7)
09 Mar 2011 Aaron Wetzler

Please give instructions on how to run. Code doesnt run nor do fig files.

19 Apr 2011 Yang Li

The file works very well. Just run stereovision.m it will come up with a GUI.

06 May 2011 sas

good

13 May 2011 Vishnu

This program shows a disparity map but if i want to save the disparity map to a file how can I do that ?

05 Jul 2011 Hossein Babaee  
19 Aug 2011 David Provencher

There seems to be an error in the integral image technique. The filtering window should be

h = zeros(WS+2,WS+2,'double');
h(1,1) = 1; h(1,WS-1) = -1;
h(WS-1,1) = -1; h(WS-1,WS-1) = 1;

that is, for a 3x3 summing window, the filtering window must be 5x5 to implement the algorithm correclty (see http://en.wikipedia.org/wiki/Summed_area_table#The_algorithm and note that points A, B and C are outside the summing window)

I tested this with a simple matrix for validation. Otherwise, thanks for the useful code.

22 Mar 2012 Yiting  
Please login to add a comment or rating.
Updates
27 Sep 2010

Adjusted description

Tag Activity for this File
Tag Applied By Date/Time
stereo Wim Abbeloos 23 Aug 2010 11:05:17
matching Wim Abbeloos 23 Aug 2010 11:05:17
computer vision Wim Abbeloos 23 Aug 2010 11:05:17
machine vision Wim Abbeloos 23 Aug 2010 11:05:17
image processing Wim Abbeloos 23 Aug 2010 11:05:17
reconstruction Wim Abbeloos 23 Aug 2010 11:05:17
3d Wim Abbeloos 23 Aug 2010 11:05:17
gui Wim Abbeloos 23 Aug 2010 11:05:17
algorithm Wim Abbeloos 23 Aug 2010 17:45:05
disparity Wim Abbeloos 08 Sep 2010 04:16:25
images Wim Abbeloos 20 Sep 2010 16:50:11
match Wim Abbeloos 20 Sep 2010 16:50:11
computer vision Rizwan 09 Apr 2011 06:32:21
algorithm alex 04 May 2011 20:54:16
stereo Olivier Ripoll 18 May 2011 09:03:51
reconstruction Charles 19 May 2011 11:39:30
disparity Fer 15 Jun 2011 09:44:42
integral image David Provencher 19 Aug 2011 11:46:40
3d Bobby Geary 27 Mar 2012 04:49:37
algorithm Sisay Shimelis 08 Apr 2012 22:42:57

Contact us at files@mathworks.com