Skip to content

CAUTION! This implementation returns inaccurate results on noisy data from real images albeit working perfectly on synthetic data. It is purposed to help understand the algorithm but not encouraged to be used in large projects..

License

Notifications You must be signed in to change notification settings

fredzzhang/Normalized-Eight-Point-Algorithm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

29 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

8-point-algorithm-MATLAB

Code Structure

eightPoint.m

Normalized 8-point algorithm implemented in MATLAB

Usage: E = eightPoint(p, q, K1, K2)
	where:
		E - essential matrix between the image pair
		p - coordinates of matched points in the first image
		q - coordinates of matched points in the second image
		K1 - intrinsic matrix of the camera corresponding to the first view
		K2 - intrinsic matrix for the second view

Note: To compute fundamental matrix, initialize the intrinsic matrix as an identity 3x3 matrix

test_synth

Test of normalized 8-point algorithm in MATLAB

test_epilines

Test of normalized 8-point algorithm on images from KITTI sequence. Epipolar lines of the given image pair will be overlayed onto the images. Below is an example.

Reference

[1] R. Hartley, "In defense of the eight-point algorithm", IEEE Transactions on Pattern Analysis and 
    Machine Intelligence, vol. 19, no. 6, pp. 580-593, 1997.

About

CAUTION! This implementation returns inaccurate results on noisy data from real images albeit working perfectly on synthetic data. It is purposed to help understand the algorithm but not encouraged to be used in large projects..

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages