Is it possible to fuse a 2d lidar (rplidar-a1) with a usb monocular camera (RGB only)?
Hi everyone, I’m trying to calibrate a 2d lidar (rplidar-a1) with a usb monocular camera (RGB only)? Is this possible or do you need a 3d lidar or a RGB-D camera in order to perform calibration between the two to achieve sensor fusion? I'm using ROS noetic and a raspberry pi 4B. Many thanks in advance
Asked by CrazyFrog77 on 2022-07-11 10:50:57 UTC
Comments
its not entirely clear to me what you're trying to accomplish.. what state information are you "fusing"? are you running object detection or visual slam using the camera? the overlap between a 2D lidar and an RGB image would be a line of pixels (or more likely due to the angular resolution of each a dotted line of pixels).
If you have a feature detection in the RGB image that corresponds with that overlap, then yes, its presumably possible to use the LIDAR information to fuse or register the detection with the pointcloud or scan.
if the goal is just camera calibration, there are usually ways that don't involve direct depth measurement (such as printing out a checkerboard)
Asked by shonigmann on 2022-07-11 13:51:52 UTC
@shonigmann thanks for the reply. The goal is obstacle detection using sensor fusion, what you're saying completely makes sense. I'm wondering if navigation stack allows for a monocular camera to be integrated with 2d lidar when creating local costmaps. Something I need to explore a lot more!
Asked by CrazyFrog77 on 2022-07-12 10:42:05 UTC