vrabaud/sfm_toolbox

Bundle adjustment parameters input

preethamam opened this issue · 0 comments

I am trying to use your toolbox that uses Sparse Bundle Adjustment (SBA) as one of the steps to optimize the motion and structure jointly. Where I need to find 3n+11m parameters for points and cameras.

Since I did not find a good reference to use your toolbox functions. I am writing this message to ask few questions.

  1. If I have 2D images for image stitching and knows the matching and their homography between the image. Do I need to triangulate to get the 3D points? So this will be the input to the bundleAdjustment function.

  2. Also if I take camera intrinsics matrix K into the consideration. How to optimize the total of 11 parameters? Which includes both the R,t, and K parameters. Because the function requires the input of the K matrix. In some cases, if the camera is not calibrated we do not know the K matrix.

Any of your inputs are much appreciated. Please let me know if you have any questions.