Occassional thoughts about orienteering
Sunday, December 04, 2011
Mapping runnability with lidarI've been experimenting with trying to map vegetation - the runnability of the forest - from laser scan/lidar data.
The map below shows one of my tests. I created the map using OL Laser. I used the software to measure object heights. I created different colors to correspond to object heights. The yellow is ground height. In this particular map, the yellow areas are mostly open fields or open water. Then I created three other categories of object height. The lowest objects are shaded dark green. Medium height objects are light green. The heighest objects are white. I figured this shading scheme would roughly match runnability.
The map below shows highlights three different sections of forest.
Section A is an area that has a lot of low objects, with some yellow areas mixed in. This area turns out to be dense cedar trees. It would be mapped dark green on the final orienteering map.
Section B is an area with mainly tall objects. In this case, the tall objects are mature trees. The area has little undergrowth and would be mapped as white forest.
Section C is an area that is mixed. It doesn't have as many short (dark green) objects as A and it doesn't have as many tall (white) objects as B. This area would probably be a shade of green on the final orienteering map.
The tricky thing about mapping the runnability of the forest around here is that the changes are generally subtle. The shift from white woods to light green woods isn't very distinct and deciding where the mark the change is difficult for the mapper. The laser scan data might be a way to help see some of those distinctions.
I plan to do some more experimentation.
Back to okansas.blogspot.com. posted by Michael | 9:01 AM
Volumetric analysis, where you bin the off-ground signal within certain heights (like from 0.5 - 2 meters) in varying bin sizes, may be more useful than simply height analysis. You're really best off starting with an existing O'map of that area or one very similar vegetatively, and then play with the bin sizes until you get results approximating the existing map.
I'm not exactly sure I know what "volumetric analysis" means. What I did was set ditfferent categories for the object heights and gave those different categories different colors. I don't have my notes handy, but the yellow was he ground + something like 0.5 meters. The dark green was something like 0-3 meters; light green was 3-7; and white was 7-15.Post a Comment