Flood mapping and analysis platform based on open satellite data and free and open source geospatial

Video thumbnail (Frame 0) Video thumbnail (Frame 1799) Video thumbnail (Frame 3729) Video thumbnail (Frame 4516) Video thumbnail (Frame 5367) Video thumbnail (Frame 12236) Video thumbnail (Frame 13933) Video thumbnail (Frame 15115) Video thumbnail (Frame 15675) Video thumbnail (Frame 16798) Video thumbnail (Frame 20439) Video thumbnail (Frame 22950) Video thumbnail (Frame 24531) Video thumbnail (Frame 25267) Video thumbnail (Frame 26195) Video thumbnail (Frame 26816) Video thumbnail (Frame 27778) Video thumbnail (Frame 28622) Video thumbnail (Frame 29432) Video thumbnail (Frame 36691) Video thumbnail (Frame 37378) Video thumbnail (Frame 37876)
Video in TIB AV-Portal: Flood mapping and analysis platform based on open satellite data and free and open source geospatial

Formal Metadata

Flood mapping and analysis platform based on open satellite data and free and open source geospatial
Title of Series
Part Number
Number of Parts
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

Content Metadata

Subject Area
Flooding remains the most widely distributed natural hazard in Europe, causing significant economic and social impact. Nowadays, availability of earth observation data generates fundamental contributions towards mitigation of detrimental effects of extreme floods. The technological advance allows development of online services able to process high volumes of satellite data without the need of dedicated desktop software licenses. The talk presents the data, the algorithms and the technologies used to develop such an online system that can use multi-scale satellite data, together with reference and in-situ information, to map the areas affected by floods and giving the users the possibility to inspect, process, analyze and validate the information. The platform, created by National Meteorological Administration of Romania, offers services based on Open Geospatial Consortium standards for data retrieval (WMS, WCS, WFS) and server-side processing (WPS, WCPS). The services were built using open source solutions such as GeoServer, OpenLayers, PostGIS, GDAL, rasdaman.
Keywords Romanian National Meteorological Administration
Satellite Freeware Mapping Texture mapping Execution unit Open source 3 (number) Denial-of-service attack Set (mathematics) Denial-of-service attack Mathematical analysis Connectionism Very-large-scale integration Computer animation Logic Software Computing platform Software framework Form (programming)
Scale (map) Slide rule Rational number Open source Mapping Decision theory Field (computer science) Denial-of-service attack Public domain Mereology Event horizon Number Data management Component-based software engineering Computer animation Operator (mathematics) Relief
Area Denial-of-service attack Denial-of-service attack Mereology Number Information retrieval Data management Arithmetic mean Data management Computer animation Information retrieval Hazard (2005 film) Quicksort
Satellite Dynamical system Context awareness Thermodynamischer Prozess Texture mapping Multiplication sign View (database) Decision theory Archaeological field survey Water vapor Real-time operating system Mereology Proper map Data management Digital photography Information retrieval Geometry Web service Mathematics Component-based software engineering Human migration Meeting/Interview Software framework Office suite Website Area Service (economics) Satellite Mapping Real number Product (business) Data management Curvature Phase transition Duality (mathematics) Chain Software framework Freeware Volume Spacetime Point (geometry) Classical physics Dataflow Service (economics) Open source Process (computing) Computer-generated imagery Image processing Denial-of-service attack Mathematical analysis Event horizon Rule of inference Theory Product (business) Number Sound effect Prime ideal Revision control Operator (mathematics) Software Traffic reporting Task (computing) Context awareness Information Surface Projective plane Mathematical analysis Denial-of-service attack Computer animation Software Extreme programming Personal digital assistant Web service Information retrieval Universe (mathematics) Revision control Local ring
Computer program Service (economics) Process (computing) Computer-generated imagery Image processing Denial-of-service attack Similarity (geometry) Raw image format Mereology Event horizon Field (computer science) Number Product (business) Chain Web service Electronic meeting system Core dump Repository (publishing) Series (mathematics) Traffic reporting Physical system Area Focus (optics) Satellite Image processing Water vapor Sound effect System call Product (business) Data management Computer animation Auditory masking Personal digital assistant Information retrieval Internet service provider Freeware
Satellite Medical imaging Computer animation Information retrieval Letterpress printing Moving average Event horizon Field (computer science) Product (business) Neuroinformatik
Satellite Standard deviation Texture mapping Multiplication sign Client (computing) Mereology Proper map Variable (mathematics) Web service Component-based software engineering Series (mathematics) Social class Thermodynamischer Prozess Computer font Satellite Kolmogorov complexity Sound effect Measurement Product (business) Digital signal processing Type theory Curvature Condition number Right angle Prototype Physical system Resultant Server (computing) Service (economics) Observational study Process (computing) Algorithm Image processing Denial-of-service attack Control flow Event horizon Product (business) Revision control Prototype Latent heat Authorization Form (programming) Condition number Standard deviation Validity (statistics) Water vapor Denial-of-service attack Sphere Power (physics) Component-based software engineering Computer animation Personal digital assistant Web service Information retrieval Case modding Revision control 5 (number)
Service (economics) Observational study State of matter Multiplication sign Computer-generated imagery Denial-of-service attack Database Price index Water vapor Machine vision Area Number Neuroinformatik Database normalization Medical imaging Goodness of fit Iteration Linear map Area User interface Service (economics) Mathematical analysis Water vapor Denial-of-service attack Database Price index Logic synthesis Computer animation Personal digital assistant Information retrieval Case modding Point cloud Musical ensemble
User interface Mobile app Standard deviation Process (computing) Server (computing) Water vapor Stress (mechanics) Price index Client (computing) Thresholding (image processing) Event horizon Number Computer animation Personal digital assistant Case modding output Musical ensemble output Physical system
Thermodynamischer Prozess Service (economics) Process (computing) View (database) Multiplication sign Image resolution Client (computing) Web browser Price index Grass (card game) Architecture Type theory Medical imaging Computer animation Software Integrated development environment Computing platform Website Reference data Resultant Row (database) World Wide Web Consortium
Area Point (geometry) Thermodynamischer Prozess Pixel Forcing (mathematics) Mathematical analysis Content (media) Interactive television Denial-of-service attack Line (geometry) Limit (category theory) Mereology Subject indexing Computer animation Integrated development environment Lattice (group) Personal digital assistant Series (mathematics) Information retrieval Physical system
Computer animation INTEGRAL Internetworking Web service Image resolution ACID 5 (number) Prototype Asynchronous Transfer Mode
Point (geometry) Metre Slide rule State observer 12 (number) Dynamical system Presentation of a group Image resolution Direction (geometry) Water vapor Distance Machine vision Number Neuroinformatik Medical imaging Preprocessor Inclusion map Meeting/Interview Term (mathematics) String (computer science) Authorization Energy level Software framework Lie group Endliche Modelltheorie Area Noise (electronics) Simulation File format Mathematical analysis Computer simulation Denial-of-service attack Measurement Wave Kernel (computing) Computer animation Ring (mathematics) Integrated development environment Quicksort Reference data Reading (process) Resultant Spacetime
Computer animation Meeting/Interview Office suite Event horizon
Computer animation
the let's start morning
everyone and I'm really happy to see that so many survived to validate yesterday good to see you here and that we have here FIR diverse talk by connectionist and he's from the Romanian National Amateur logical agency is gonna talk about framework and form flood mapping data using their wits all partners he will about it you know was that is not only the person represents here he is a very important member of always due community and helping out so much that you should surely by beer if you catch around somewhere so what do you think you understand how this very easy to catch because if you see the speed of traveling right now after the gun the unit is not so so here go on is full of dangers and now I might have some problems my foot so would apologize to you but and I would like to that on the chair so please don't then give set of me OK so full danger her
as you all know what we witnesses they um an avalanche of data where this is coming from all kind of sources and the some of that data has a has dual components still geo is geographic data and can be used in a number of domains and in the last year I was involved in my institution in in activities related risk management disaster prevention and and mapping of this work was this disasters so but I will be talking about flights today because in my country as you could see it highlighted in the and and on the slide how I'm really has a problem with with with flooding so we had a major flood events in and in 2002 thousand 5 thousand 6 2008 2010 2015 14 and the other from a lot of things central was affected farmlands or inundated and unfortunately many many people or or or killed by this by this let's say we try to do our part 2 to help the 1 involved in in managing the crisis and and relief operations yeah Madison
flaws means and is how Romania look sort of Romanian floods let we map in the last 16 years look on the on the on the map so you could see that quite important part of the country was affected by by by flooding and some of the the the the roof things you're seeing there sometimes a flood happen in more it using the same in the same area so it's quite quite a problem and if we go to plot
management solutions there are a number of things that could be done for flat proficient flood
management but not like this this is a famous victory my country a see the prime minister of former prime minister and 2 ministers caring in a boat moving in the flooded village uh um just giving like instructions to people with we don't really doing proper crisis management and this as we have also a shame about how how the government and the fact that the
flat and so what they can do is to use the given the better I have is the use of surface originate J to map the phylogeny to to do like to the dynamics of the flight and to offer some kind of a had value-added information to the 1 involving in and In the operation so in this context we did we developed satellite-based flood mapping service and its it's covering all the phases of the of the of the crisis so starting with the pre-crisis only do risk analysis maps and classical risk crimes map uh of course the most important part of this services is during the crisis when do we we want to map the the the the affected areas by floods and we do damage assessment and we try to do this this in near real time 1 of you with the with the background in in the 1st of decision who know that's not not like in movies you could not really get set like the top but theory of time has passed sometime in the tonsil so have thousand all of problems but we try to do it to have a look at the map and at least in in 24 hours or fastest possible uh and then you also using survey data to see how their how did the affected areas are recovering after the flood events of event infrastructure and agriculture feel that and so so that we use also is both optical and radar data to to map the floods and we try to to list such as possible with and without too much human intervene intervention we started its service in 2005 the new version was released in 2007 and 2012 was important but year for us because we the migrated from proprietary software to foster G we already start to do that in 2007 not parts of our infrastructural or based on a free and open source software but since 2012 restore totally yeah continue using the free software and a new version was started in late last year and the component of that with this service will be presented to you we do this work in the framework for a number of research projects on on pension that we're not doing just past the met office but we have some part of Romania space agency and actually sees of hydrology and water management universities involving algorithm the development and and and and so on and our beneficiaries are everyone because what you do is public but that the user we targeting the general Inspectorate for Emergency Situation Ministry of Environment and local councils and preference when it went with this kind of events happening entirely defined kind of kind of work before you all flows of products strictly delivering case of of crisis so we have a dual near-real-time from maps this is not to time but it's like some hours were made a day or maybe 2 days after the flood event and have not finance which is quite the same but it is validated and somehow and it may be looks better described as from the from the graphic point of view we do matching flood extend maps flat area classification this is as when you see a flat map usually you something they need didn't rule and you know it is what but the weather is not as deep everywhere so we try to have a more elaborate product that it was what was defined some classes of shallow what they're what they're interested in vegetation 1 with this deep water we do damage assessment maps reports and we do animations because the floods usually involves and what is moving so animation as corticaw or way to to to to make understand how they would be the intervals and free stuff that the chain this is presented to the image processing change presented here so we needed a data tasks a prior for satellite data uh our services that we don't have we have some support and prices but you don't really have money for for
data so we rely mostly on the free data available and that's not something he stop and sometimes we get free data from commercial companies just because a crisis and that they they could help you but you're not paying 108 . so to data is 1 1 events happen we select the best data available for for for that area and we do a number of steps for image processing make corrections referencing MOS 18 and then uh then we extract the and the what the mask and the creator of cartographic products to show to show the affected areas and very fast to try to and ideation by comparing with some other imagine you're going on the field the GPS receivers and forced 4 in the last years we are also working closely to with similar services but it it with the international scope of maybe you know the European Commission is running the Copernicus program and 1 of the core series of Copernicus is in emergency management map management systems and
their full of our focal point in Romania we we actually the trigger this service when it's when it's when it's needed and another partner with work with in the inner part of the international traffic which were sex that the beach similar effect of Copernicus service um get in a couple of cases we received several data for the coming traffic tragedy and we were the the the service provider that if this college creating our products and then we create thematic products like maps with the flights we do damage assessment reports and hasn't call animation and someone from the products and looking like
this so classical problem with the the events on the but on some kind of background would be a map or satellite image and we have some explanation because this maps usually are getting Printed and the that the people who using 1 field or maybe the computer it was designed to be a printable maps this kind of
this have shown more examples of this that has a mention we also
later the flies and who could also Stuart happens in time you see is an example from the the Danube flat in 2006 had a huge flood it it it lasted for plants and the effect it has some very large parts of of land and it took some time is and dikes break 1 by 1 and and and the world of some villages were really heavy flooded 1 here was completely destroyed into which really moved after that by the authorities so we do this kind of things and and then but this
is nothing new this is something that many other series of doing what is new in the last version of the service we launch if you try to empower people so to offer up right proper form of data processing so you create your own flood maps and do some analytics on on the on the on the results so for that we added some new components client and server to be able to process satellite data remotely without having to download anything when you compute computer and reduce we reticent standard like WPS and CPS and them well but it is not going to have only a prototype so we created a prototype and a case study using not all the event in 2006 because we have a lot of data for that event and it it it was really a large event but the idea is to move forward so for that event had only Morty's daytime and use after and that such data for validation purposes and you some more details about and why would do something like that of course firefighter but it's going to end with a boat in the and and see if people will not need to create his own flat that's of course for for this kind of clients or users but the class services is is what is needed and this is to continue with that some other type of of users like researchers of the but will try to understand what happened during the flood after we Europe to adjust to to big better measures so 4 of these guys is important maybe to have a tool that could allow that to do that we without the specific knowledge of image processing and but 1 or more of the reason is that the fact we have this kind of or a sphere using to map the flat but it's is like that of global but there where we think this kind of uh refer to work in all parts of Romania but sometimes you have to look condition very different than and each of which are not always perform brighten when you if you interesting to have like the most accurate products and you could you could pay yourself with the with the data and then we can get the best we try to do something fast and reliable enough which is not perfect and this is you could you could do your own if you have local knowledge it's it's it's it's far far better OK so as I mentioned have this
case study on the the Danube and in the affected areas from the by the flooding in 2002 discuss it could see highlighted under on map for this areas of the
band some some data on it David the time on this image is formed from April to December so a database we also drive a number of indices from from this data which explain later from from these images of course there are also affected by clouds of you don't have each day good state it's OK and the data was published fooled used combine services services is was put available through a web interface for users and also we need some to play with some of some other kind of data to to do analysis and we choose to work with and he II and I'll show you what to do with that and it took 15 years over the centuries of about 8 days and he I so I mentioned before from from more these data we extracted some value and it's data so Ford floods the number of indices that could help you there to discriminate the water from the images so the number of them from the we extract the best 1 from this scientific papers and we we try to put the preprocessed this images were for the priority that these so the 1st 4 our this kind of in indices and the last 1 is is something different this is so from computer vision is the image annotation technique but
so as I mentioned grant number of this kind of methods to to to discriminate water but they have the pluses and minuses and the idea was to be able to to threshold the the bands and this kind of approach compared to select perfectly to the wall so but this scenario for from is the
user is using the the web interface is a date from that event the spent has input respect to take up the case not quality and to to to query the data this is then it selects the imports of them back more his band on the indices and set thresholds for each of them he he trigger the did the processor WP is about but that the CPS with some rough abroad happened in behind and until the weather must would be displayed on their under the client the the use of and could adjust the threshold or and ask for additional statistics and and uh and play that they're the powerful
bitterly as I mentioned open-source software what is important is that we use we have all these services for view process and download the results we used for data processing recent uh and across the grass demand G OTB the web client is created with openEAR seen some other well-known libraries and do we store record the time chassis that's that's a classical this yeah for this kind of
things we could also point made by just applying but not the processing is this is not done by by day in the desert environment the web site is very easy looking something like that we have some data and on the left reference data and prepaid predefined indices you could expect this and you could trigger the the time Navigation select the type of process you want to do and it just into the servant because it's always databases very fossil is that the amount of data is this high but as the size of the images in sense of that because the data resolution of and
then you could do some kind of analysis so you could define some areas and requiring the system to show it in this case we used in EVI to see if I can some areas you how DVI evolve in in different years so you can see here you have what happened from 2004 thousand 14 so that the the flood forces was 2006 and you could you could see or it could take on 1 point
on a map on a pixel lattice flooded because see how the motion of the and yet and of a vegetation index and explain so you could see let the red line In 2006 it 6 you but it was affected and was able to recover after I don't know free in the mountains something and could do this on any other what parts of the limits of interaction without knowing how this process behind could save data and they with that in your own environment after that if your content so
we also provide some data for violations of of for a few days we have hydrogen a high-resolution Austria and lots of data so is very good because usually speaking the same 2nd as the the the the mode data with high resolution we been preprocessed Oster when you could do your own using the data we provided could could compare with what we have an acid see how how it how it performs and usually spike good but you could you could get a little more and that's it I
would try to move to operational we try to that they all of this is still too weakly to Internet sense that they we just use more because we had this strike a flat but it we week where doing it suddenly integration went ecstatic they did I discovered it and and and so on I thank you and given the
fact that could not don't ask me to hard question because you think you
correct space for questions that the you thank you I was at 1st very comprehensive presentation of some things that are to ask what more although we have done already I haven't done the 1st is that you know you are probably month is that while the framework directive on water classification taken that into our account then it's the other but with the environment people something mitigation processes like restoring wetlands and using their also all of that's in the main points so after this slide in 2006 uh the authorities start thinking about for activation of the Americas in the in the in the past this was the dynamic has to extend no flooding the court at the level those villages in the past but the dike would constructed for agricultural purposes and this happens after 2 weeks of heavy heavy water is device simply collapse and now we have this discussion of the ring operating the areas and this is a tool for them to understand how the flood happened and to put some more to have some reference data to to play with and you don't want to make a point to decide to on and destroying the dice and that the dance again freely yeah so that our partners account of the needs of the uh yes so much of was Freeman directed at work do you have a much better chance but what about is the no it's it's it's it's it's a big believer and the Danube no small it was in that area but in the middle of a river somewhere we have to modify because of pre-industrial something as we were having modified this is no I think you will see that they are from questions they just said that data models that addiction and that analysis models are not perfect enough and get through the snow model in computer vision that has 100 per cent in distance so the question is which fire or which are the issues that you presented this data to noise preprocessing of images not perfect this will never be perfect the more he says there is a land your resolution so horses 250 meters of 1st 2 bonds and have 100 meters the ideas but they you could have data data that's that the temporary resolution is very good so you have to yeah it's it's an advantage that this advantage but if you if you provide such a pool where you could say that you know you have water and that is the the in lies between this value and this value and in the end of that number 1 of them should be between this and this and that you could construct pretty a pretty good model to fit the format Extract correctly that the what even using was modesty reading just help hopes and and it'll performed was much higher resolution but then the temporary solution not so good what the them have fewer incorporated some kind of simulation in your what do you think about it our partners from the hybrid judge did some simulation actually actually in terms they validated their models using this in this kind of results but is not directly into integrated itself it would be nice if we had to take some measurements public some of building a dam or something that you could to calculate how a new flood wave foot would go through it now and yellow dots and if you have that was known for his work the research work but my point is claim observation is I don't know if you're looking at an eternity initiative of using a kernel framework that fits techniques among the upper strings To conclude amount of water that you get the people less potential to attract OK and not but if you look at the scale of thank you thanks for the sort of related to the question about the inclusion of simulations to
use in a forecaster triggered a consistent or not so this a restricted when usually most trouble with don't be prepared in advance because and I'm from the Met Office but don't really use the forecast is not so fast when this kind of events happen it has for and and then some even happened yeah where 1 it might happen but it's not like activation is not greater than the
forecast that is alright thank this is generic the questions you wish to attention he will be very slow today so things in the past