See 'Monocular_odometry.py' This is an example of visual localization based on mono camera sensor. Intrinsic parameters (K matrix is known) .
- Detect feature descriptors (harris corner, SIFT)
- Track the features in 2 consecutive images, using optical flow.
- Estimate Essential matrix (E) between points correspondences, using RANSAC model fitting.
- Estimate camera rotation (R) and translation (t) based on essential matrix.
- Update the current pose based on R and t.
See 'Stereo_odometry.py' Camera intrinsic and extrinsic parameters are known
- Undistort the images 2.Detect features in Left_img_tk, Right_img_tk, Left_img_tk+1, Right_img_tk+1.
- Estimate features and track them across 4 images
- Compute the sparse disparity map and estimate the 3D location of the points at time t and t+1
- In the FrontEnd, estimate the initial transformation between the pointclouds at time t and t+1 using the Iterative closest point cloud (ICP) method.
- Construct a PoseGraph and optimize the poses.
- Plot the result
Bundle Adjustment example, to optimize camera poses, 3D points locations and intrinsic parameters. See 'BA.py'