How to triangulate two 3d lines

Click For Summary
SUMMARY

This discussion focuses on triangulating the 3D coordinates of an object using two cameras in a computer vision project. The line equations derived from the cameras are represented as 1: a1x1+b1y1+c1=0 and 2: a2x2+b2y2+c2=0, with known parameters a1, a2, b1, b2, c1, and c2. The Fundamental matrix (3x3) is utilized to establish the relationship between the two cameras. Since the lines do not intersect, the solution involves calculating the shortest line connecting both lines and determining its midpoint as the object's center.

PREREQUISITES
  • Understanding of 3D coordinate systems
  • Familiarity with line equations in geometry
  • Knowledge of the Fundamental matrix in computer vision
  • Proficiency in vector mathematics and transformations
NEXT STEPS
  • Study the derivation and application of the Fundamental matrix in stereo vision
  • Learn about calculating the shortest distance between skew lines in 3D space
  • Explore the use of change of basis matrices in 3D transformations
  • Investigate algorithms for triangulation in computer vision projects
USEFUL FOR

This discussion is beneficial for computer vision engineers, 3D graphics developers, and anyone involved in projects requiring object localization using multiple camera systems.

assafMOCAP
Messages
2
Reaction score
0
Hi guys,
I am working on an computer vision project.
the project uses two cameras to triangulate an object in front of the cameras.


Homework Statement


Express the object's location in 3d coordinates relative to the cameras.

Homework Equations


From the software i can get two line equations.
The line equations are for lines going through the centers of the lens and the center of the object itself
1: a1x1+b1y1+c1=0
2: a2x2+b2y2+c2=0
Known parameters are the a1,a2,b1,b2,c1,c2
Also know is the relation between the two cameras (The Fundamental matrix 3X3)
Also, the object coordinates in 2d on a projection plane from each camera is known (i.e. x1,y1 and x2,y2 )

Sadly, the lines do not meet, so its also needed to calculate the shortest line that connects both lines and treat its middle as the object's center.

The Attempt at a Solution


how to calculate a vector that originates from the middle of the two cameras, to the object.
 
Physics news on Phys.org
assafMOCAP said:

Homework Statement


Express the object's location in 3d coordinates relative to the cameras.

Let's say that your original coordinate system is ##(O,\vec i,\vec j,\vec k) ##
You want to set a camera in ##O'## and ##O''##, and set a coordinate system in ##O'## and ##O''##.

If ##M## has coordinates ##(x,y,z)## in the original coordinate system, its coordinates ##(x',y',z')## in ##(O',\vec I,\vec J,\vec K) ## satisfy

## \begin{pmatrix} x \\ y \\ z \end{pmatrix} = \vec{OO'} + P . \begin{pmatrix} x' \\ y' \\ z' \end{pmatrix} ##

where ##P## is the change of basis matrix from ##(\vec i,\vec j,\vec k)## to ##(\vec I,\vec J,\vec K) ##. Similarly for the other camera.
 

Similar threads

Replies
17
Views
3K
Replies
1
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
5K
  • · Replies 11 ·
Replies
11
Views
3K
Replies
11
Views
5K
  • · Replies 1 ·
Replies
1
Views
7K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
1
Views
2K