Some EO sources have a wide, high-res
PANchromatic band, together with the multispectral sets.
So we can pan-sharpen: downscale each low res layer
using Brovey, Baysian, Gram Schmidt, PCA or
now SRGAN for those who like generative neural nets .
#gischat #limrigis #resolution
Add some lines to #LimriGIS,
Rules of the game are hard to miss.
Merge your words in limrick format,
Any topic from cartography to spatial stats.
Don’t mince your words, don’t be remiss.
We met a Doctor from Dakar
who follows climate change along Sahel’s corridor.
To test for time series change in rainfall
we checked for trends with Mann Kendall,
confirming decreased rain from negative tau and small p.val.
#gischat #limrigis #rainfall #climatechange #sahel
WGS84, ETRS89, OSGB36
Why are datums so complex?
Some believe the earth is flat,
so they can skip the question off the bat.
But GISers need to fix a CRS for all their tricks.
#gischat #limrigis #crs #flatearth #wgs84
Making global maps today’s a breeze.
Choosing data is picking cherries from the trees.
If it’s elevation that I need, SRTM is quite fine,
and average annual climate data from BioClim.
Cities, boundaries, population, open geodata, all for free.
#gischat #limrigis #srtm #geodata #bioclim
I spoke to Zoe from south of Sydney
who wanted to rotate her map strategy.
She tried to orient south at the top.
Why not, it’s arbitrary. Give it a pop.
While most choose north up, there’s no such decree.
#gischat #limrigis #orientation #gis #northarrow
In GIS we sometimes try resample
to get a higher resolution scramble,
extract all info from each pixel,
scrape each bit of onion from the grittle.
But will the egg and omlette still resemble?
#gischat #limrigis #resample #resolution
I knew a guy from the Isle of Wight
who had a shapefile that wasn’t quite right.
We chose to convert to geopackage,
and thus get rid of sidecar baggage.
But cleaning up topology then caused an outright fright.
#gischat #limrigis #geopackage #Shapefile #topology
Today we have remote sensed data in petabytes
from hoards of sun synchronous satellites.
With spectral bands combined in indicies
and many aggregations of time series,
who know how much research has not be tried.
We mappers see the world in vectors, grids or CRS converts.
Our messages use visuals, with words we’re rather terse.
But if your ear is tuned to rhyming,
with a cadence and good timing,
Go design a layout using 5 short lines of verse.
#gischat #limrigis #vector #grid #crs #limrick
When overlaying polygons is part of the maneuver,
the question “intersects or intersection” begs conjecture.
If I need a polygon layer to clip the other,
then st_intersection works just like a cookie cutter.
While st_intersects returns a list of overlapping features.
#gischat #limrigis #intersection #spatial
A mapper up in Bern was showing urban transportation.
All streets were digitized, ignoring any junction.
She looked for handy operations
to split each line at intersections,
and found that GRASS v.clean does just that function.
#gischat #grassgis #limrigis #intersection
The square root of 2.0, squared is flippant.
Should be = 2, unless I’m unintellgent.
In math, for sure, but computers use floating point format,
So decimals with base and mantissa are always inaccurate.
The result depends on bits in the significand.
> sqrt(2.0) ^ 2 == 2
[1] FALSE
Thanks to:
https://cran.r-project.org/doc/FAQ/R-FAQ.html#Why-doesn_0027t-R-think-these-numbers-are-equal_003f
https://stat.ethz.ch/Teaching/maechler/R/useR_2014/Maechler-2014-pr.pdf (slide 29)