Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>This problem is much more 3d related than image processing.</p> <p>What you are trying to do is to register the knowing 3d and since for all the frames there is same 3d points->camera relation the transformations calculated from registration will be the camera motion transformations.</p> <p>In order to solve this you can use PCL. It is opencv's sister project for 3d related tasks. <a href="http://www.pointclouds.org/documentation/tutorials/template_alignment.php#template-alignment" rel="nofollow">http://www.pointclouds.org/documentation/tutorials/template_alignment.php#template-alignment</a> This is a good tutorial on point cloud alignments. </p> <p>Basically it goes like this:</p> <p>For each pair of sequential frames 3d point correspondences are known, so you can use the SVD method implemented in</p> <p><a href="http://docs.pointclouds.org/trunk/classpcl_1_1registration_1_1_transformation_estimation_s_v_d.html" rel="nofollow">http://docs.pointclouds.org/trunk/classpcl_1_1registration_1_1_transformation_estimation_s_v_d.html</a></p> <p>You should have at least 3 corresponding points.</p> <p>You can follow the tutorial or implement your own ransac algorithm. This will give you only some rough estimation of the transformation (can be quite good if the noise is not too big) in order to get the accurate transfomation you should apply ICP algorithm using the guess transformation calculated at the previous step. ICP is described here:</p> <p><a href="http://www.pointclouds.org/documentation/tutorials/iterative_closest_point.php#iterative-closest-point" rel="nofollow">http://www.pointclouds.org/documentation/tutorials/iterative_closest_point.php#iterative-closest-point</a></p> <p>These two steps should give you an accurate estimation of the transformation between frames.</p> <p>So you should do pairwise registration incrementally - registering first pair of frames get the transformation from first frame to the second 1->2. Register the second with third (2->3) and then append the 1->2 transformation to the 2->3 and so on. This way you will get the transformations in the global coordinate system where the first frame is the origin.</p>
    singulars
    1. This table or related slice is empty.
    1. POCalculating camera motion out of corresponding 3d point sets
      singulars
      1. PTQuestion
    1. PTAnswer
    1. This table or related slice is empty.
    1. USMichael Kupchick
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. POCalculating camera motion out of corresponding 3d point sets
      singulars
      1. PTQuestion
    1. This table or related slice is empty.
    1. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTBountyClose
    3. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTAcceptedByOriginator
    1. COThank you for that answer! I will check how it's done with the PCL library. But as far as I see it, where is the difference between ICP algorithm and the one I used in OpenCV? I mean sure both work differently, but don't both of them return the relative rotation and translation between the frames? The tutorial gives me a 4x4 matrix T. I've tried it just the same with the data I get from the openCV function. I also created a 4x4 matrix P, which is the orientation and position. I tried to calculate the trajectory by: P_t = P_t-1 * T_t Unfortunately this gives me impossible results. Any ideas?
      singulars
      1. PO
      1. USValentino Cazalet
    2. COTry to check the estimateAffine3D function. Generate a set of 3D points then generate the second one with known translation (no rotation). Try to register those sets and see if the results are reasonable.
      singulars
      1. PO
      1. USMichael Kupchick
    3. COThanks! I added an update to my post above that shows two simple test cases for the estimateAffine3D()... it seems to work for one test case but not for the other... so I guess I gotta try to use the PCL library instead... Or do you have any other ideas?
      singulars
      1. PO
      1. USValentino Cazalet
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload