Matera - Day 2

After a surprisingly swift intro to Linux and bash the day will evolve around the GDAL library as well as the - relatively obscure - pktools.

Session 1 - gdal

EPSG / Spatial Referece Information

• gdalinfo does not include the nodata value in the statistics, mind the data-type

• gdal_translate can be used for tiling (as an example) -projwin ulx uly lrx lry

• when cropping with gdal_translate (or any software) the cropping window does eventually not match perfectly, so the software will resample (or shift) automatically in the background

“The VRT driver is a format driver for GDAL that allows a virtual GDAL dataset to be composed from other GDAL datasets with repositioning, and algorithms potentially applied as well as various kinds of metadata altered or added.”

• VRT can be used to link together many files or create subsets (extents) of images

  gdal_translate –formats | grep ENVI - find if gdal supports the format you want

• ogrinfo can have very useful ouput an also supports SQL statements

  ogrinfo -al shape.shp ... OGRFeature(poly_fr_10poly):0 id (Integer64) = 2 region (Integer) = 2 POLYGON ((3872295.18072289 2681195.78313253,3915993.97590361 2666629.51807229,3901427.71084337 2615647.59036145,3872295.18072289 2681195.78313253))

Session 2 - gdal, bash scripting

• OpenEV - part of FWTools

  openev TCmean01-10_1km.tif - quickly view a image or vector (uses GDAL) STRL + \ - kill application

• more AWK stuff

  for file in *.tif ; do gdalinfo -mm $file | grep "Min=" | awk -F "," '{ print $2 }' done

• set comma separation, print second column

• get resolution of each TIF file and display only those with 240 pixel

  for file in *.tif ; do echo $file $(gdalinfo -mm $file | grep "Size is ") done | grep "240"

• with added basename function:

  for file in *.tif ; do echo $(basename $file ) $(gdalinfo -mm $file | grep "Size is ") done | grep "240"

• Bash string manipulation

  grep -v "inverted" - do an inverted grep

Session 3 - pktools

pktools are based on gdal but go further in many ways; e.g. extract the bounding box coords without grepping or awking. Written in C++, good documentation, relatively narrow focus.

Session 4 - openforis oft-tools

• quite usable tools from Open Foris

• learning strategy: 1) GDAL 2) PKtools 3) Open Foris tools

• oft-stat can be used for zonal statistics (very fast):

  oft-stat -i INPUT.tif -o output.txt -um INPUT_MASK.tif >> INPUT_MASK = rasterized vector file

• oft-clump and oft-bb can be used for preprocessing by finding the correct bounding boxes to do the zonal statistics with

• todo: do a benchmark between QGIS, pktools and GRASS (maybe SAGA, perrygeo gist, Python package, starspan, R, R - multicore, PostGIS, GeoTools) zonal statistics and oft-stat (be aware of the zonal freaks)