Use of the withheld bit flag

[jdnimetz]

I'm not sure this is a LAS WG issue, but, as this does relate directly to the LAS file format, I thought I'd post here for feedback.
The topic I'd like to discuss is proper use of the withheld bit flag. My suggestion is that the withheld bit flag be assigned for all points that are determined to be geometrically unreliable or cannot be reasonably interpreted as valid surface returns. Common examples of invalid, non-surface points include aerosol back scatter, laser multi-path, airborne objects, and sensor anomalies. The withheld flag may be used in conjunction with other classification codes (e.g. low/high noise), but it should be used in all cases where the previously mentioned criteria are met.

The main reason I am suggesting this: Any routine relying on good data such as density assessment, relative accuracy assessment, point cloud classification, DEM creation, feature extraction, etc. should be coded to automatically ignore the withheld bit flag in all cases. Another reason- taking an 'either/or' approach to using the withheld flag or noise classes is confusing and inconsistent.

Thoughts from this group? Is this an appropriate discussion for the LWG or perhaps better represented elsewhere such as in the USGS 3DEP LBS?

Josh

[milenajaniec]

I support Josh's idea - think it would be a good topic for discussion during the next meeting. From personal experience it seems that utilization of withheld flag is interpreted differently by the USGS partners. Perhaps we could look into whether this should be better clarified/defined?

[esilvia]

As a reminder to everyone from our meeting in March, I agree that the withheld flag is appropriate to use when points are generally undesirable for an assortment of reasons, including the ones Josh listed, plus points that are geometrically unreliable. That said, I don't believe that it is up to the LWG to prescribe that all "noise" classes always be flagged as Withheld. If USGS chooses to make that a requirement in the LBS , however, I believe that is their prerogative.

I intend to reserve a chunk of time on the next LWG call (5/18) to cover this topic and the topic of Overlap/Overrage (#5 )