Create disparity map / calibrate cameras

gmfricke commented 4 years ago

Please take a look at how cameras could be used

opportunity for ML?
Rviz
Gazebo camera position and view

Carter90 commented 4 years ago

Also take a look at image_view rosrun image_view stereo_view stereo:=/scout_1/camera/ image:=image_raw _approximate_sync:=True _queue_size:=10

Carter90 commented 4 years ago

Looks like we will have to create our own map https://gitlab.com/scheducation/srcp2-competitors/-/issues/141

Carter90 commented 4 years ago

For openCV here is what I had in the swarmie https://github.com/BCLab-UNM/SwarmathonIV-Cabrillo-Public/commit/34349ef8547d1a6b109b82419cab8cb6c292548f

Carter90 commented 4 years ago

Tracking issue on https://github.com/BCLab-UNM/SC2/tree/calibrate_cameras

Antonio-Griego commented 4 years ago

@htasnim and I have written a node that generates a disparity image. Also, however, there is a standard ROS package that does this and also provides some other functionality:

http://wiki.ros.org/stereo_image_proc

Based on what we've learned, it seems a viable step would be to integrate this existing ROS package into this project, and continue to write/integrate other packages that process that data.

For example, if we are doing SLAM there is the gmapping package:

http://wiki.ros.org/gmapping

If we wanted to go the machine learning route, and needed image segmentation, there is an old package we could use as a base to start with:

http://wiki.ros.org/pal_vision_segmentation (please note it is no longer maintained and would have to be updated and/or rewritten)

Carter90 commented 3 years ago

Needs calibration

BCLab-UNM / SC2

Create disparity map / calibrate cameras #6