1

I produced a DSM raster from LIDAR data for eastern Kentucky. There are strange lines from southwest to northeast. Do you have any idea what those might be? I generated the DSM raster with both first return and non ground classes but results were similar. I am using ArcMap 10.6. Basically I used first return or non ground classes and created DSM from elevation data by using LAS dataset to raster tool. Then subtracted DEM from it. DEM looks fine. I also created Canopy Density Model (CDM) by using LAS Point statistics as raster (points above surface/ all points). I still have the same issue with CDM. Red areas show zero value (ground) green areas are higher areas with structures or canopy. Here is the picture

enter image description here

  • @AndreSilva I believe it is not overlapping but more like no data acquisition problem as red areas display the ground. – Amadeus Sep 5 '18 at 20:45
  • Can you share a sample size (point cloud) from a small part where the problem occurs? Can you add more details about how you used the ArcGIS tools? I mean which parameters did you use in each of them? – Andre Silva Sep 5 '18 at 22:34
1

They may be an artifact of imperfect calibration during (or after) data acquisition. The stripes look like they might be the field of view of the sensor as taken along a flight path. Possibly the thin, more reddish, lines are where these scans overlap.

If this is the case then simple post-processing options might include: a median filter and applying an offset. A fancier (but not necessarily better) approach might be to try and filter out frequencies in an FFT then inverting back to raster.

  • I believe it is related to imperfect calibration probably during data acquisition. Thin reddish lines where pixels have zero value. I guess they could not collect data from those areas somehow. I tried focal statistics (I can not use median with floating data) mean and sum results did not change. What is FFT? How can I apply median filter with offset? can you please explain further? – Amadeus Sep 5 '18 at 20:41
  • 1
    It could also be the difference of vertical datum applied to the flight line though the red toward the ephemera seems to indicate a sweep (oscillating) scanner mixed in with data from a rotary scanner. There also could be imperfect ground toward the end of the strip as most of the returns are in the canopy resulting in a lower portion of ground strikes. Your overlap ground could be in class 12 (ASPRS overlap class versions up to 1.3) so perhaps have a look for that class and see if your data is hiding there. – Michael Stimson Sep 5 '18 at 23:30
0

After taking a closer look to raster and lidar data, I figured greener (higher) data acquired when trees have leaves attached while more red data most likely collected in winter or very late fall. I concluded like this because there is no problem with structures or conifers. When the majority of land type is forest and leaves do not exist. Most of first return collected from ground and eventually marked with zero height value. Most likely, those red stripes are where the overlaps are which is caused by too many ground points. Thank you for your comments and help.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.