Skip to content

Latest commit

 

History

History
68 lines (48 loc) · 2.51 KB

README.md

File metadata and controls

68 lines (48 loc) · 2.51 KB

pose estimation

For one base image we should find it in another image and draw a 2d pose of object

1-find keypoints with FAST algorithm or SIFT
2-calculating descriptor with ORB/BRIEF algorithm
3-matching keypoints with FLANN algorithm
4-calculate transformation matrix with RANSAC: R,T
5-transform corner of object to second image

base image

second image

matching

matching

For one base image we should find all objects in another image and draw a 2d pose of object

1-find keypoints with FAST algorithm or SIFT
2-calculating descriptor with ORB/BRIEF algorithm
3-matching keypoints with FLANN algorithm
4-calculate transformation matrix(projective) with RANSAC: R,T
5-transform corner of object to second image
6-mask that area and repeat 1-5

base image

second image

matching 1

matching 2 after mask

pose

For one base 3d image we should find 3d pose of object in another image and draw a 3d pose of object

1-find keypoints with FAST algorithm or SIFT
2-calculating descriptor with ORB/BRIEF algorithm
3-matching keypoints with FLANN algorithm
4-calculate transformation matrix from 3d to 2d with PnP with RANSAC: R,T
5-transform corner of object to second image
6-use Linear Kalman Filter for bad poses rejection

base image

second image

3d pose