I am somewhat confused on the meaning of the
Overlap flag which is present in las spec 1.4.
The ASPRS LAS Version 1.4 Spec states:
New to LAS 1.4 is the fourth flag, OVERLAP. It is frequently desired to thin data in the overlap region of lidar flight lines (or, in the case of mobile, "takes"). This operation is used to either improve the statistical accuracy of the lidar data or to simply thin the multiple density data in these regions. Using a Class to indicate overlap (the technique prior to LAS 1.4) caused problems in that one loses the actual classification of the point (e.g.,Ground). This leads to problems in analyzing accuracy in the overlap region. What exactly comprises points that should be designated as Overlap will be determined (as it is now) by the procurer of the lidar data. For example, USGS has an extensive description of how points should be designated as Overlap.
lasoverage readme states:
Below an explaination based on what Karl Heidemann of the USGS once told me regarding "overlap" versus "overage" points:
"Overlap" applies to any point in one flightline or swath that is anywhere within the boundary of another flightline. Whether or not a point is overlap is determined, basically, at the time of flight. It is a fundamentally immutable characteristic.
"Overage" applies to any point not in the "tenderloin" of the flightline or swath, that is, the central core of the flightline that - when combined with the others - creates a non-overlapped and gapless coverage of the surface. Whether or not a point is overage is determined by the operator (or software) that defines the single-coverage tenderloins of the flightlines. Hence, it is a subjective characteristic -- it depends solely on where somebody (here: lasoverage.exe) chooses to draw the boundary.
Unfortunately the LAS specification uses the word "overlap" where it should be using the word "overage".
It seems the LAS Spec is using the overlap flag to refer to "overage" as defined in the lastools readme since the "overlap" points are subjectively defined by the "procurer" of the lidar data. So, how should overlap points be handled? Should they be removed or used?