Chris,
Thanks for the pointer to CF, and the suggestions regarding my data.
I think the following [1] conforms to your suggestion to use 1-d arrays.
I’ll work towards adopting CF as my iterations progress.
For now, I’m happy to reinvent the vocabulary since my focus is to expose the
provenance of OPeNDAP requests of NetCDF files.
Regards,
Tim
======
netcdf disneyland_lidar {
dimensions:
point = 5 ;
variables:
float northing(point) ;
northing:units = "UTM meters; zone 11" ;
float easting(point) ;
easting:units = "UTM meters; zone 11" ;
float altitude(point) ;
altitude:units = "?" ;
float intensity(point) ;
intensity:units = "?" ;
int category(point) ;
data:
northing = 414783.24, 414782.09, 414782.73, 414782.03, 414781.51 ;
easting = 3741669.72, 3741669.31, 3741670.04, 3741670.63, 3741671.03 ;
altitude = 40.32, 40.44, 40.40, 40.28, 40.34 ;
intensity = 26, 3, 10, 26, 22 ;
category = 9, 9, 9, 9, 9 ;
}
======
[1]
https://github.com/tetherless-world/opendap/blob/570c638afb05c6397fc60606b9eaf7ff36fd917c/data/source/usgs-gov/earthexplorer/version/disneyland-2014-Jan-31/manual/CA_OrangeCo_2011_000402.sample.cdl
On Feb 3, 2014, at 12:26 PM, Chris Barker <chris.barker@xxxxxxxx> wrote:
> On Mon, Feb 3, 2014 at 6:09 AM, Timothy Lebo <lebot@xxxxxxx> wrote:
> How can (or should) LiDAR be modeled in NetCDF-4?
>
> NetCDF is very flexible -- if you want to know how best to represent
> particular data in a standard way, you'll want o use a defined standard. Most
> commonly used is the "CF" metadata standard (CF for Climate Forecast, but is
> has been expanded and adopted for all sorts of other things):
>
> http://cf-pcmdi.llnl.gov/
>
> A quick search of the standard name table does not turn up "lidar", but I'd
> be really surprised if it hadn't been used for that, so I'd puruse teh
> standard and see what matches your data well.
>
> I’m a NetCDF novice, and it seems to me that the format assumes uniformly
> spaced samples (e.g. [1]).
>
> NetCDDF works really well for regular gridded data, but it by no means is
> restricted to that. You certainly dont want to arbitrarily shift your data
> points to a regular grid.
>
> Below is a table of sample points that I’d like to encode in NetCDF. It
> originates from a USGS LAS file about Disneyland.
>
> lat lon alt intensity
> category(Ground=2;Low-Vegetation=3;Unclassified=1;Water=9)
> 414783.24 3741669.72 40.32 26 9
> 414782.09 3741669.31 40.44 3 9
> 414782.73 3741670.04 40.40 10 9
> 414782.03 3741670.63 40.28 26 9
> 414781.51 3741671.03 40.34 22 9
>
> if I"m reading this right, you have a whole bunch of points, each with a
> lat-lon, and a couple values and a flag associated with that. If there is no
> particular structure to the arrangement of the point, then you want to keep
> it as unstructured data:
>
> a bunch of 1-d arrays, length-number-of-points long:
>
> latitude
> longitude
> alt
> intensity
> catagory
>
> it's pretty much that simple, but browse through the CF standard to see how
> you want to label then and what mets-data to apply so that users will be most
> easily able to interpret the data.
>
> I think you want to look at what CF calls "Discrete Sampling Geometries". If
> it's unclear what do do a post to the CF list would probably result in good
> advice:
>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
> My latest attempt is at [3] and included below, but it isn’t satisfactory
> because I’m forced to round my points to suit a grid.
>
> yeah, you really don't want to do that...
>
> -Chris
>
> --
>
> Christopher Barker, Ph.D.
> Oceanographer
>
> Emergency Response Division
> NOAA/NOS/OR&R (206) 526-6959 voice
> 7600 Sand Point Way NE (206) 526-6329 fax
> Seattle, WA 98115 (206) 526-6317 main reception
>
> Chris.Barker@xxxxxxxx
