GRASS GIS 7: your reliable geospatial number cruncher

Video in TIB AV-Portal: GRASS GIS 7: your reliable geospatial number cruncher

Formal Metadata

GRASS GIS 7: your reliable geospatial number cruncher
Title of Series
CC Attribution 3.0 Germany:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Release Date
Open Source Geospatial Foundation (OSGeo)
Production Year
Production Place
Portland, Oregon, United States of America

Content Metadata

Subject Area
GRASS GIS (Geographic Resources Analysis Support System) looks back to the longest development history in the FOSS4G community. Having been available for 30 years, a lot of innovation has been put into the new GRASS GIS 7 release. After six years of development it offers a lot of new functionality, e.g. enhanced vector network analysis, voxel processing, a completely new engine for massive time series management, an animation tool for raster and vector map time series, a new graphic image classification tool, a "map swiper" for interactive maps comparison, and major improvements for massive data analysis (see also The development was driven by the rapidly increasing demand for robust and modern free analysis tools, especially in terms of massive spatial data processing and processing on high-performance computing systems. With respect to GRASS GIS 6.4 more than 10,000 source code changes have since been made.GRASS GIS 7 provides a new powerful Python interface that allows users to easily create new applications that are powerful and efficient. The topological vector library has been improved in terms of accuracy, processing speed, and support for large files. Furthermore, projections of planets other than Earth are now supported as well. Many modules have been significantly optimized in terms of speed even by orders of magnitude. The presentation will showcase the new features along with real-world examples and the integration with QGIS, gvSIG CE, R statistics, and the ZOO WPS engine.
Keywords GRASS GIS geospatial analysis HPC cluster image processing vector raster voxel time series visualization Python WPS
Goodness of fit Graph (mathematics) Information Software developer Mass Grass (card game) Number
Area Metre User interface Distribution (mathematics) Pixel Functional (mathematics) Histogram Open source Combinational logic Bit Mass Line (geometry) Grass (card game) Mereology Computer Revision control Software Different (Kate Ryan album) Profil (magazine) Endliche Modelltheorie
Scripting language Point (geometry) Stapeldatei Game controller Software developer Planning Bit Grass (card game) Word Vector space Term (mathematics) Representation (politics) Endliche Modelltheorie
Area Game controller Group action Texture mapping Mapping Digitizing View (database) Multiplication sign Drop (liquid) Line (geometry) Grass (card game) Computer 2 (number) Vector space Different (Kate Ryan album) Network topology Boundary value problem Software testing Mathematical optimization
Point (geometry) Classical physics Functional (mathematics) Algorithm Graph (mathematics) Multiplication sign Graph (mathematics) 1 (number) Bit Database Travelling salesman problem Distance Graphical user interface Centralizer and normalizer Software Different (Kate Ryan album) Order (biology)
Point (geometry) Laptop Divisor Computer file Multiplication sign Mereology Semiconductor memory Different (Kate Ryan album) Computer hardware Operating system System programming Endliche Modelltheorie Mathematical optimization Modem Curve Multiplication Surface Bit Grass (card game) Software Vector space Personal digital assistant Calculation MiniDisc Window Spacetime
Laptop Principal ideal Point (geometry) Satellite Statistics Principal component analysis Modulare Programmierung 2 (number)
Point (geometry) Web page Functional (mathematics) Computer file Open source State of matter Software developer Online help Streaming media Grass (card game) Wiki Programmer (hardware) Computer programming Operator (mathematics) Order (biology) Flowchart Energy level Social class Library (computing)
Slide rule Functional (mathematics) Process (computing) Computer file Different (Kate Ryan album) Search engine (computing) System programming Shape (magazine) Open set Line (geometry) Computer architecture
Standard deviation Algorithm Functional (mathematics) Mapping Computer file INTEGRAL Interface (computing) Sound effect Shape (magazine) Menu (computing) Grass (card game) Word Process (computing) Computer programming Operator (mathematics) Entropie <Informationstheorie> Normal (geometry) Endliche Modelltheorie Booting Resultant
Statistics Process (computing) Graph (mathematics) Code INTEGRAL Line integral Operator (mathematics) Database Grass (card game) Social class
Scripting language Algorithm Multiplication Service (economics) Mapping Spline (mathematics) Variety (linguistics) INTEGRAL Correspondence (mathematics) Zoom lens Image processing Set (mathematics) Maxima and minima Parameter (computer programming) Grass (card game) Web 2.0 Message passing Process (computing) Doubling the cube Different (Kate Ryan album) Arrow of time Feature space Error message Descriptive statistics
Dependent and independent variables Digital photography Interface (computing) Order (biology) Energy level Endliche Modelltheorie Graph coloring Thresholding (image processing)
Satellite Scripting language Virtual machine Cluster analysis Cloud computing Staff (military) Drop (liquid) Grass (card game) Power (physics) Supercomputer Wiki Process (computing) Different (Kate Ryan album) Core dump Operating system System programming
Point (geometry) State observer Functional (mathematics) Texture mapping Information Multiplication sign Workstation <Musikinstrument> Electronic mailing list Set (mathematics) Volume (thermodynamics) Database Line (geometry) Wiki Frequency Volume Visualization (computer graphics) Vector space Metrologie Analytic continuation Spacetime
Pixel Average Personal digital assistant Multiplication sign Mathematical analysis Content (media) Computer Position operator
Area View (database) Self-organization Time series God Spacetime
Area Time zone Data management Personal digital assistant Profil (magazine) Cuboid Denial-of-service attack Volume (thermodynamics)
Point (geometry) Video projector Open source Multiplication sign Source code Combinational logic Parameter (computer programming) Black box Open set Information technology consulting Revision control Mathematics Different (Kate Ryan album) Authorization Software testing Computing platform Stability theory Control system Area Time zone Mapping Line (geometry) Grass (card game) Software maintenance System call Shooting method Software Procedural programming
Point (geometry) Software Multiplication sign Sampling (statistics) Website Freeware Vulnerability (computing)
so good morning everybody I'm happy to see you here for the grass years 7 talk college you're reliable number cruncher um the idea is to showcase what has happened in the past so uh almost 6 years since we started graph 7 development initially very carefully but then with the increasing speed so I even having a lot of information here want to consider it as a kind of lightning talk where we try to get to school of a massive amount of news so I I really regularly
receives that's people have some idea what grass yes might be how could have been something like this so these are screenshots from the eighties which also indicates that the software exist for now 30 years which is something pretty exciting it's probably 1 of the oldest the software packages out there in the open source geospatial arena but it is continuously developed and this is that big difference to some other pictures so what has been happening is of course some evolutionary and you can already guess that uh we got in your user interface competing new user interface many new functions and features which I will walk you through now yeah there's no something harmless uh what you find also its where histogram towards this uh more slides what those in the room would already have been using grass in the past maybe the version rests 4 5 6 I got to know even across to user in the past days and maybe also here so nothing special you want to draw your medical profile and of course you can do so what's already interesting that you can do this also with massive amount of data so we are testing for example showing uh 48 billion pixels that is the current 25 meter below elevation model of Europe you can do so and not also for packages are able to do that you can also combine legend and histogram you see over there uh the the distribution in the legend something which is already a bit more fancy and pretty useful especially if you are in this area here where we have very homogeneous values and then only some differences and part of the
OK then you can be do the obvious thing draw lines around lines your that declines whatever something which reminds you of that geodesic support is also not found in all those jails all their but it requires some special computation and something you still
probably a bit in development let's say I'm not all nice features are which we want to see there about something pretty promising in new geospatial model this enables you to all create a graphical workflow and as a bonus you can write it although the Python script so at this point you can graphically combine your steps into a workflow and then eventually turn this into a script maybe to further modify modifications in future but especially run this as a batch job so it is easy to go from a graphical representation to something automatic
the a few highlights in terms of uh vector data processing uh grass yes is a topological GIS it has always been a topological 1 and there's really no plan to change this to superfeatures because we think to maintain quality unitary the topological control and the possibility to apply topological truly a proper logical words to your data um along with that we have
found digitizer topological digit as of course this enables you to get uh directly control over your newly created data you can see if they are topologically correct something if you're not familiar with that just to give you an idea if you have adjacent areas the shared boundary is 1 Mondrian not true boundaries as engine but simple features this makes a difference Rico difference because you know if you are not precisely digitizing the and you will be surprised tested people digitizing nowadays this is something which is still relevant of course then once you have 1 shot boundary you cannot have gets in natural and this is something which renders the idea interesting you see from the future all of you there and that we also have full support fallback Porter Baker vector drop map so that you can have an underlying map from which you want to digitize or you can even copy features from that if it is a Osterman those sorry a vector map and a 6 so the especially
interesting about the topological opaque and in Graz GS is that it became much faster this is something what you usually have more quality at the expense of more computation and this is this holds true for a vector topology as well you have to do more competitions because you have to check if it is correct or not and but you can see here comparing grass 6 grass 7 the grass 7 line is the group of uh the green line which is almost horizontal of course it is not perfectly horizontally and that's possible but you see the increase of uh seconds to do some computation is really dramatically lower than it was for cross 6 and turn while officially grass 6 is the stable view or you can already get of course the the the grass 7 snapshots in time and we have been making tests for which uh Victor met with huge vector maps to understand and where optimization can done
another topic is vector network analysis a rich tool set available at it now comes also with a graphical interface you can see from the uh will you can easily select your various algorithms you want to run a shortest path of the classical ones are there but there are also other britches visibility network and or centrality and other to it's about which are probably a bit different are giving you some extra features so what is next vector network analysis for if you're not familiar with that you take a graph which could be a street network and then you want all and moved on top of this this is a classical example year and traveling salesman problem you have several points to visit on your road network and you want to understand what is the optimal past the between these different points which is not necessarily the shortest distance but you could also use attributes and say OK traffic plays a role you can make it dynamic and fetch the dynamically traffic data from some database and then run this thing on top of so you would even get different graphs in function of the time of the day the switching to
arrest that data support for massive frosted data this is something which we have been working on in the past 2 years in order to be able to the yet to be fast with also a huge datasets is you
know data are growing like crazy and hardware probably less also at this point we need optimization of software but the question is what means massive NSF is something which is not so easy to define of course the size of the dataset plays a major role but it is also related to hardware resources software capabilities and operating system capabilities the cross mighty uh it's a part of its system which runs on multiple operating systems Windows Macintosh in Oaks uh a IIx whatever not uh various systems and at this point it really depends on the operating system you're using limiting factors as we have already heard yesterday in a in a talk from uh is not that costly anymore but but still way more costly than disk space and so in many modem commands in grass affair find the opportunity to switch from a ROM-based memory model to a small based disk-based memory model in case that your memory is not sufficient you are able to all uh outsourced kind of computational to notice this takes more time naturally but still you can do it even on limited of systems like a laptop for for example the largest supports then largest supported file size is also unusual and we have been working on implementing even if all the vector data the possibility to exceed on 32 bit systems um the very often if you become like also here a nice curve to show this is the cost surface calculation and you see from a nonlinear increasing computational time has been changed to a linear 1 which makes quite some difference of course you have million of points in the 2nd so compute the thing um I have here is a snow
of example of sorry to see this is a
laptop here you can perform the CA on this good that is the principal
component analysis with 30 million points which you'll get from a satellite data for example in only 6 seconds and this is something which you cannot easily all in some other statistical software packages and so what else do we have with a
lot of new tools for hydrology and you see here a kind of flow charts a stream towards a general order a segmentation basins what else is there in uh that whatever you can get yourself there's some scientific publications are available and especially in the grass wiki you can find a date pages dedicated to that in order to see how to compute things and on again we have been testing uh the watershed or hydrological tools with enormous amount of data and at this point we always happy to receive comments my file still bigger and it can't be done and then we can see if we can do something about it but it starts to become difficult to find such a state then
what it's too can programming you can uh performed on the development of course as before it is open source but there some help for this and to something completely new is the Python API so we have now Python support integrated and so this is something which already gained interest in the mating is we see new users popping up which we have never seen before and they say OK I'm doing Python programming in Python programming is fairly easy and so the grass API gives you now have the possibility to connect not only to the commands themselves but you can also connect to the underlying functionality even at low level it means that the C library level and that just makes everything together as you need it again but we have further the programmers manual which is documenting something and with a new by grass courts uh API and integrated in the normal manual where you can find classes to do topological operations and so on you can
also use and process a bitch system you don't even have to all started for using it but you can just put some Python lines together and then do a complete batch-processing starting from your shape file or you to file or whatever you have to retrieve something out of it and then put somewhere into a different system or online with WP as whatever you need publication open access is describing the architecture that and I guess the slides will become available so that I would make mine available for sure so you can go through and check that programmers many
already mentioned you have an i search engine there and the full functionality of a document OK this leads since we
speak about programming that you can use crosses bake own on the consumer
side to this has already been done for a while so this integration in Q GIS for the programming the former 6 standard and this allows us to all them it's a year should years user but still in use of the grass functionality effect you just doesn't provide what you need and it works like this that you have in the main menu the uh processing tools that here there are also other 2 was reduced like saga out of the mood to words and so forth and in some way industry if you have crossing stored you also find your grass commands and if you if you want to and calculate for example watershed you looked at elevation model GeoTiff whatever into 2 GS go there from watershed and you get a watershed out and and it will be shown again in and the normal Judea GIS interface which means that internally um the grasses called everything is calculated and then uh the uh the result returned is due for shape from this is another example of the disorder algorithm based on uh not to boot so here we can solve this map to something else that operation ownership file and you get a cationic this the
so even more complex integration again with processing you can load to adjust the random workflow to illustrate what's possible uh you can fit your data from post from opposed years database or from W as as a web-service or WMS whatever into Q GS you go to processing do operation and get lecturers and this is something which is really coming up nicely in the DNA code sprints we have been updating this processing uh sets to August 7 so you have both in parallel our grass 6 supporting precedent for the
furthermore we have our integration this is something which existed to 2 thousand but just to remind you that this is possible their spatial classes in are available and additionally this sp grass 6 it is to cut 6 but it runs equally with grass 7 and like this you can go into your grass sorry into your our session uh and fetch data from the graph database and reduce your statistical analysis on top of the this is something which we use a lot in our research and
eventually uh the WPA support Web Processing Service uh with integration in a for what to put the 50 to North Lobo there uh into or double pi WPS and 250 to north like this you have the possibility to create on web processing services using the almost 400 uh commands which are available in grass and what's particularly interesting you can see here uh this exam style of documentation that is the set expressed in and say to so to say the explanation of each grass command so you take a grass command and you query it what is your WPS process description and it will return this description and like this you don't have to manually reduced all the various parameters of a grass command but it is just passed from this 5 so and extra bonus is here that you if you write your own scripts Python shared whatever you prefer using the grass pass even your own script will do the same thing again because this is generated by the cross passes so this is applicable to all comments so quick walk through image processing with annual geocoding toward here's an example you have some unregistered historical map there loaded some OpenStreetMap and this enables me to find the different ground control points corresponding once it tells you what the arrows you can do error minimization and then run a variety of we transform algorithm like the polynomic or lands course and have signed available
for image processing of multiple channels with the possibility to evoke scatterplots now you can of course a zoom into them you can look into your your feature space there you can also perform
classification supervised classification as it was already always there but with a new interface to that means you lost your in Europe for example RGB composite or sum a false color composite you can digitize inside you get the spectral response all day and then you can use that to train your classification model furthermore with unsupervised
classification this is also a new and this allows you to do a segmentation based classification can see here using different thresholds you can decide on which level you want to segment the thing and from an order photo tool something like that you can use either segment so to
say something about the the bigger data we have set of possible the possibility to all ground grass on supercomputers clusters and so forth uh as mentioned already with the possibility to all compile grass on more or less any operating system and these BCR come sometimes which with the unused for me unusual operating systems but it doesn't matter and we are able to not only run grass on a single core but you can make use of the job system the Q-system which is commonly used on supercomputers because it is a shared resource or if you want to do cloud computing you can fire up your virtual machines and then uh just run your staff remote so in our uh mean my research foundation we have been using Grid Engine for that like this i you write a small script to launch the different jobs for example uh satellite data reconstruction or something like that and the drop engine with then distribute of all these notes there uh the drops to compute them in power that's not much effort really in the grass wiki you can find
documentation and eventually some mentioning of the new possibility for temporary uh support we have space-time tupes now available in grass yes so new
space-time functionality here and you can see there's already a rich set of commands in the 1st place to define the your container like uh spaces space-time jewel which means you say OK my dataset starts in this year and ends in that year and semantically we have monthly observations or hourly or daily or whatever and then you put in in a simple list of the names of the fights there it can be rest vector or a volume of volumetric trust them and along with the time step and then it will automatically register everything properly and you can also do get filling for example recalculate missing maps and so forth again also here of a scientific publication available and into wiki page just to give you an idea US visualization tool was the timeline tool which shows you all in in the 1st line of for example a point point in time data which are for example metrological observations which appeared regularly from the station and the other 2 are continuous or not continuous information as sorry data
covering the period for example a month or a year or whatever it is this you can all you define and then since it is registered in this uh database you can then do
aggregation an easy way which means if you want to hold the calculates the annual summer temperature you just define your someone say OK give me the average and that's it that's 1 come and likewise you can do or a
climate change analysis takes 32 years of data and then do the uh aggregation so here we have computation here we have some example this is a plot of a single pixel in a stake of data and you get out in this case chlorophyll content over time in 1 position on you we can also flies through our data this is a timeline of modus a LST data on land surface temperature which we have in the
so this leads us to all these organizations this is
small as the last topic i want to show here and your animation to its are also available and as you can already guess from here there you have it not only into the what we have seen before but also in 3 D uh which means you can define your view of this is an example for like God they uh light uh data time series moving due on which we have been showing 2 days ago and the the workshop space-time to workshop here down to find the material online and this enables you to a gets really an idea of what's going on in your area when you have time series available
for that to the case where additionally annual swiping toward this is interesting for disaster management for after this is the tsunami in Japan as an example you see uh how the flooding zones changed and you have this wiper and you can just check uh the inner detailed mold zooming into or what has happened in that area what do these ization toward
also enables you to show volumes and volumes are in this case and created by small boxes that are volumetric pixels and start to look into a volume you need to use transparency for example or a isosurfaces or pro fights and then you can put your profiles into your volume like uh being seen although there and more you can use transparency to get an idea of what's going on there and eventually the
possibilities since we combine big data and vizualization there's some nice nice kind of theater at North Carolina State University and you can just project your shoots sets they're onto something uh with this a combination of different video projectors and this enables you to really uh get a better idea on Teutsch data something especially tricky is this due area coastal zone here because it's extremely long but they're not very high at this point but it is not easy to look into a mnemonic to about if you're lucky to have something around the corner and you can do a vizualization like that so to
summarize the we propose grass as before a platform for sustainable open science but of course also for a consultancy or whatever you are you doing the keywords are here reproduced the stability we can reproduce things because we have the source code available this applies to any open source software and for science this is far as their natural habitat uh because otherwise in the black box research side we can figure out what really have to or what other authors have been doing some return of investment is another keyword um size in example we have commands which have been developed long time ago and they're still available but if you now on top of that developed for example uh uh some procedures you can so you are still able to do that to use those even 10 or 20 years later this is something exciting maybe not for the younger people hear about those who were doing GS stuff for a long time they know what you remember I've done something something like 10 years ago that cracked the script and see what happens and we're pretty sure that with minor modifications you can even run them in gross 7 and we have documented the changes from the previous version is the parameter has been renamed to something more reasonable you find it in about 10 documentation you have history preserve each map set which is the uh workspace that preserves its history for ever and reliability we think that uh the new testing and quality control system which I haven't shown here but you can find it on line is something which is the way to go to figure out our if everything continues to work and longevity for open science the court integrated in uh grassed yes survives even longer means if you have a contribution to the these contribute because we try to add that we have also grass add-ons which is fairly interesting and this gets a kind offer gratis maintenance in future because we say OK if it is on the grass infrastructure and we have some changes internally we just propagated through all the call so that is something which may interest
you OK I'm done with my time where to get the stuff gross website with weak mating used what as documentation free sample data which also using the examples this enables you to all play around with the software if you didn't do so and uh at this point I would like to conclude thank you very much


  343 ms - page object


AV-Portal 3.20.1 (bea96f1033d39fbe77f82542458e108105398441)