search

tysik / obstacle_detector

A ROS package for 2D obstacle detection based on laser range data.
540 stars 205 forks source link

Calibration of two Lasers #16

Closed ov-vo closed 5 years ago

ov-vo commented 5 years ago

Thanks for your package. It works for a Laser (SICK S300). There are two Lasers that I don't know the exact relative position. Do you have any idea about the calibration of two Lasers?

tysik commented 5 years ago

Hullo, That's a very good question. I was working on a calibration script back in the days but didn't eventually finish it since once prepared manual corrections worked for me very well. Anyway you can do that (I assume that both scanners have parallel scanning planes - we only work on correction of yaw angle in 2D plane):

  1. Place your testbed next to a wall that you know is straight.
  2. Run Rviz and display two separate laser scans.
  3. Look for a "dissonance" between points laying on the wall.
  4. Modify the TF of one of laser frames iteratively so that the wall points from both scan visually lay up on a line.
ov-vo commented 5 years ago

It's a good idea to calibrate according to straight wall. Thanks.