2

I'm trying to classify/remove flight line overlaps in .las-files using the Whitebox toolset. The LAS file has a band in the centre where two flight lines overlap.

When I run wbt.classify_overlap, the tool also classifies a lot of points that are not located within the overlapping area, why could this be?

If I run wbt.flightline_overlap the result shows values between 1-3 over the whole area, but with more 3's in the actual overlap zone. As far as I can tell, there isn't even 3 flight lines to begin with.

If I use the ArcGIS tool "Classify LAS Overlap", it only classifies points within the overlap zone. It would however, be more convenient in terms of licensing to be able to use Whitebox. What could be the reason for this difference? I'm specifying the same parameters on the same dataset.

0

1 Answer 1

1

The difference between the ArcGIS tool output and the WhiteboxTools output is the result of the ways in which they identify points belonging to different flightlines. The ArcGIS Classify LAS Overlap tool uses the Point Source ID property to determine each point's associated flightline. By comparison, both the WBT ClassifyOverlapPoints and the FlightlineOverlap tools identified the condition of points belonging to different flightlines based on a threshold time difference. This technique does generally work well, but unfortunately I had hard-coded the time difference threshold value, rather than making it user-specified. It's likely the case that the value used was not well conditioned for your particular data.

In fairness, the approach of using the Point Source ID to infer the flightline is more conventional. The main reason that I did not adopt this approach when I first wrote these tools was that sometimes the Point Source ID information is either omitted or lost from files. However, I have recently added the new RecoverFlightlineInfo tool, so that when this condition occurs, users have a mechanism in place to update their Point Source ID with relative flightline information. As a result, I have just committed updated versions of the two flightline overlap tools to use the Point Source ID instead, similar to the ArcGIS solution. The ArcGIS tool flags all points in a grid cell that have the same Point Source ID as the point with the maximum (absolute) scan angle. I have included this option, however, I have also added three other flagging criteria for conditions when scan angle information is unavailable or inadequate. These include:

Criterion Overlap Point Definition
max scan angle All points that share the PSID of the point with the maximum absolute scan angle
not min point source ID All points with a different PSID to that of the point with the lowest PSID
not min time All points with a different PSID to that of the point with the minimum GPS time
multiple point source IDs All points in grid cells with multiple PSIDs, i.e. all overlap points.

I have also updated the help documentation for these two tools so that they better describe the operational characteristics of the new method. These new versions of the tools will be available in the next public release of the WhiteboxTools open-core, which I am planning sometime towards the end of April or beginning of May. If you cannot wait until then, you may either build from source code, or contact me for an pre-release binary, which I will be happy to provide.

2
  • Hey! Thank you for a great answer! I figured there was a difference in the method between the tools, but couldn´t really find any specifications on how either worked. I'm in no rush, I can use the ArcGIS-tool until the next WhiteboxTools release, but thank you for the kind offer.
    – Emanuelv
    Mar 28, 2022 at 13:02
  • Hello! I have a follow up question. I checked in on the ClassifyOverlapPoints in the user manual and it looks as the changes you describe above have been implemented. However, when I run the tool through Python I still can´t specify the flagging criteria. I´ve tried installing and updating through both pip and Conda. The pip installation is version 2.1.2 and the Conda forge is 2.0.0. What am I missing?
    – Emanuelv
    Jun 22, 2022 at 7:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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