Skip to content

eugeniu1994/Visual-Odometry

Repository files navigation

Monocular-Visual-Odometry

See 'Monocular_odometry.py' This is an example of visual localization based on mono camera sensor. Intrinsic parameters (K matrix is known) .

  1. Detect feature descriptors (harris corner, SIFT)
  2. Track the features in 2 consecutive images, using optical flow.
  3. Estimate Essential matrix (E) between points correspondences, using RANSAC model fitting.
  4. Estimate camera rotation (R) and translation (t) based on essential matrix.
  5. Update the current pose based on R and t.

Stereo-Visual-odometry

See 'Stereo_odometry.py' Camera intrinsic and extrinsic parameters are known

  1. Undistort the images 2.Detect features in Left_img_tk, Right_img_tk, Left_img_tk+1, Right_img_tk+1.
  2. Estimate features and track them across 4 images
  3. Compute the sparse disparity map and estimate the 3D location of the points at time t and t+1
  4. In the FrontEnd, estimate the initial transformation between the pointclouds at time t and t+1 using the Iterative closest point cloud (ICP) method.
  5. Construct a PoseGraph and optimize the poses.
  6. Plot the result

Bundle Adjustment example, to optimize camera poses, 3D points locations and intrinsic parameters. See 'BA.py'

About

Stereo & Mono Visual odometry

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages