| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
I would like to edit the map produced by a SLAM algorithm, to add in no go areas. These areas do not ave a physical barrier. Would greying them out be sufficient to prevent the A* algorithm used by AMCL from directing the robot into them?
If not, how would one accomplish this?
Also, can AMCL make use of .png maps, or does it have to be the default pgm?
Thanks
I've manually edited maps using GIMP and had great success with keeping planners from making paths through unwanted areas. I used black instead of gray as that corresponds to "very" occupied. From the wiki:
When comparing to the threshold parameters, the occupancy probability of an image pixel is computed as follows:
occ = (255 - color_avg) / 255.0, wherecolor_avgis the 8-bit value that results from averaging over all channels, e.g. if the image is 24-bit color, a pixel with the color0x0a0a0ahas a probability of 0.96, which is very occupied. The color0xeeeeeeyields 0.07, which is very unoccupied.
As black is 0x000000 it should yield 100% (or very) occupied. Gray may not result in an area being considered occupied.
Edit:
According to the wiki that I linked to, you can use a PNG instead of PGM, but PNG is not supported on OS X.
As @Procópio stated in the comments
painting the map black will have the side-effect of degrading localization, as the real LIDAR measurements will not match those of the edited map.
Asked: 2018-02-12 19:32:02 -0500
Seen: 234 times
Last updated: Feb 13 '18
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
