Go to file
Phyks c8f262bedf Cleaned the repo and added some doc 2013-08-01 19:40:47 +02:00
Articles Cleaned the repo and added some doc 2013-08-01 19:40:47 +02:00
Rapport Initial commit 2013-07-28 15:24:49 +02:00
Test images Cleaned the repo and added some doc 2013-08-01 19:40:47 +02:00
.gitignore Added a .gitignore file 2013-07-29 11:20:30 +02:00
README.md Cleaned the repo and added some doc 2013-08-01 19:40:47 +02:00
RGB2grey.c Initial commit 2013-07-28 15:24:49 +02:00
Values used for camera specs Cleaned the repo and added some doc 2013-08-01 19:40:47 +02:00
bmp.c Initial commit 2013-07-28 15:24:49 +02:00
makefile Initial commit 2013-07-28 15:24:49 +02:00
sobel.c Initial commit 2013-07-28 15:24:49 +02:00
stereo.c Updated license info 2013-07-29 11:16:36 +02:00



A program (in C) to find the distance to an object using two images and computing distances accross these images (thanks to a stereoscopy algorithm).


This program computes the mean distance between a stereoscopic camera and an object. We assume both cameras have the same specs (which are the diameter of the field stop of the cameras and the focal length).

The idea behind this software is very simple. First, it tries to determine the distance in pixels between the same element through the two photos (either by using a FFT-based approach or by comparing patterns from the two pictures). Once it gets the distance in pixels through the two pictures, it can easily convert it to a real distance between the camera and the object, using the specifications of the camera.

For more details about the way I convert a distance in pixels to a real (perpendicular) distance in centimeter, please refer to the following article (I just used the formulas in this paper without major modifications) : http://docs.lib.purdue.edu/cgi/viewcontent.cgi?article=1064&context=ecetr

This software was tested successfully (ie giving a result correct according to the incertitude on the camera specs) with a standard compact camera, taking two pictures separated by a few centimeters (to mimmic our eyes).

Notes :

  • I use the OpenCV library to easily load and go through the images. OpenCV is a C library for computer vision licensed under a BSD License ( http://opensource.org/licenses/bsd-license.php )
  • The software computes both X and Y displacements but will focus on X displacements to determine the mean distance of the object as the cameras are supposed to be on the same horizontal plane (Y displacements are vertical displacements and should be negligible). It is possible to take into account a vertical displacement with minor modification to the code (you should just find a correct formula in 3D :)
  • X corresponds to the abscissa and Y to the ordinate of the pixel in the image. They both goes from 0 to the size of your image.
  • It has been successfully tested on a GNU/Linux system. I think it could work on Windows with minor or no modifications but haven't try.

Pixels are numeroted this way :

(0,0)		(1,0)		(2,0)		...		(Width,0)
(1,0)		(1,1)		(2,1)		...		(Width,1)
 ...    	  ...		  ...		 ...		   ...
(0,Height)    ...		  ...		 ... 	 (Width, Height)


You must have opencv libs installed on your system.

Command line arguments

-l (--left)
Path to the left image
-r (--right)
Path to the right image
-D (--diameter)
(float) Diameter of the camera's field stop
-f (--focal)
(float) Focal length of the camera
-d (--delta)
(float) Distance between the two cameras
-s (--size)
[Optional] (int) Size of the subsets used, default is 50
-x (--abscissa)
[Optional] (int) Abscissa of the pixel to compute the mean distance
-y (--ordinate)
[Optional] (int) Ordinate of the pixel to compute the mean distance
-c (--choose)
[Optional] To display the image and select which pixel you want to work with (same as -x and -y but useful if you don't know the coordinates of the pixel)
-o (--output)
[Optional] Path to a file to store the output
-S (--sobel)
[Optional] Run a Sobel edges detection algorithm on both left and right images first to work only on edges. This option won't have any effect if working in single pixel mode.
-F (--fft)
[Optional] Use cross-correlation and fft based method (faster but a little bit less accurate).
-H (--hamming)
[Optional] Use an hamming window to improve FFT (but slightly slower).
-h (--help)
Display a help message


Please, send me an email if you use or modify this program, just to let me know if this program is useful to anybody or how did you improve it :) You can also send me an email to tell me how lame it is ! :)


I don't give a damn to anything you can do using this code. It would just be nice to quote where the original code comes from.


Phyks (phyks@phyks.me) wrote this file. As long as you retain this notice you
can do whatever you want with this stuff (and you can also do whatever you want
with this stuff without retaining it, but that's not cool...). If we meet some 
day, and you think this stuff is worth it, you can buy me a <del>beer</del> soda 
in return.

Note about the content of this repository

This repository contains :

  • An Articles folder containing the PDF of the articles I found on the Internet about this subject and that I intensively use to code.
  • A bmp.c code to load a bmp image (personnal implementation, not used in the definitive version - I used OpenCV to handle all formats instead).
  • A Rapport folder containing a report on the code in French. Not very useful I think...
  • A RGB2Grey.c code to converts RGB images to Grey levels images (idem it's a personnal implementation of the OpenCV equivalent. I used OpenCV in the definitive code).
  • A sobel.c code to apply a Sobel on an image (idem, personnal version, used the OpenCV equivalent in the definitive code)
  • A stereo.c code that is THE MAIN AND DEFINITIVE CODE that handle the stereoscopic measurement.
  • A Test images folder containing some images I used to test the algorithm. They have been taken by myself and you can do whatever you want with them (except those showing a table with a bottle of wine which has been taken by somebody else and for which all rights are reserved).
  • A Values used for camera specs file containing the params I used for my camera. These should be relatively standard and give a good approximation for standard cameras.