visual slam test result

asked 2020-11-29 00:14:45 -0500

dinesh gravatar image

updated 2020-11-29 00:15:34 -0500

I am testing the orb slam2 and openvslam ros package. I'm using logitech c270 monocular camera for this purpose. Till now i've done this tasks: 1. Calibrate the camera. Here are the calibration data:

image_width: 1280
image_height: 720
camera_name: camera
camera_matrix:
  rows: 3
  cols: 3
  data: [1332.107090741891, 0, 646.8646682329983, 0, 1283.553969935012, 355.9515438263148, 0, 0, 1]
distortion_model: plumb_bob
distortion_coefficients:
  rows: 1
  cols: 5
  data: [-0.6490073621987695, -0.2343410863844743, -0.02115290462789926, 0.0202207718179566, 0]
rectification_matrix:
  rows: 3
  cols: 3
  data: [1, 0, 0, 0, 1, 0, 0, 0, 1]
projection_matrix:
  rows: 3
  cols: 4
  data: [1331.066284179688, 0, 646.359274794886, 0, 0, 1281.771240234375, 355.4571682230744, 0, 0, 0, 1, 0]
  1. Run openvslam and orbslam2 slam node, create map of environment.
  2. Run the openvslam localization node and test the relocalizaton accuracy. Here is the demo video.

But as in the video also. The relocalization accuracy is very bad. i.e the robot is unable to localize itself within given map. Should i be using depth sensor for robot to localize itself accurately in given map.

edit retag flag offensive close merge delete