[pdal] Obtaining "stats" without reading points

[Howard Butler]

I would say that most of the LAS files I've seen have good bounds, but it is not uncommon to find a mismatched one. The histogram seems to be more frequently mismatched than the bounds. Once upon a time, you would also find mismatched record counts, but with LAS 1.4, those kinds of files are now unrecoverable.

libLAS (and older versions of PDAL) have had a "give me the headers whatever they are, but don't trudge through all the data" method, and the "QuickInfo" stuff that Andrew is describing which is now available in master speaks to the capability. 

Howard


> On Dec 3, 2014, at 1:55 PM, David Burken <dburken at comcast.net> wrote:
> 
> On las data I had, the scanned min, max values were exactly the same as the header.  Can't say I check a lot of data though.
> Dave
> 
> 
> On 12/3/14, 2:11 PM, Kramer, Oscar wrote:
>> Hi Andrew,
>> 
>> JSON would work. 
>> 
>> Re the untrustworthy data: that's a bummer. But perhaps certain dimensions are more trustworthy than others, and the user will know which ones are good for specific datasets. In those cases, getting to the header data is still a worthwhile effort.
>> 
>> Oscar
>> 
>> From: Andrew Bell [andrew.bell.ia at gmail.com]
>> Sent: Wednesday, December 03, 2014 11:47 AM
>> To: Kramer, Oscar
>> Cc: pdal at lists.osgeo.org
>> Subject: Re: [pdal] Obtaining "stats" without reading points
>> 
>> On Wed, Dec 3, 2014 at 9:51 AM, Kramer, Oscar <OKramer at radiantblue.com> wrote:
>> This is related to Michael's issue post here: https://github.com/PDAL/PDAL/issues/569. I need access to min/max bounds on quantities like intensity values, geographic bounds, etc (specifically in las files but should be generalized). I know those values are available in the las header. Yet it seems that to obtain these "stats", I need to employ a stats filter and read all the points. Is there another mechanism for quickly reading certain general properties (like bounds) that may be available from the header?
>> 
>> I've written the infrastructure to provide what you're asking, but I don't think I ever hooked it up to a command-line app.  I'll do this shortly.  But in the end, it doesn't matter if you can't trust the data in the header, as Michael suggests is often the case.  This really is too bad -- not much point in having the data if it's not correct.
>> 
>> We were talking about providing JSON output.  Would this work for you?
>> 
>> -- 
>> Andrew Bell
>> andrew.bell.ia at gmail.com
>> 
>> 
>> _______________________________________________
>> pdal mailing list
>> 
>> pdal at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/pdal
> 
> _______________________________________________
> pdal mailing list
> pdal at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/pdal