DigitalGlobe and Open Source

Video thumbnail (Frame 0) Video thumbnail (Frame 846) Video thumbnail (Frame 1393) Video thumbnail (Frame 2265) Video thumbnail (Frame 3721) Video thumbnail (Frame 4608) Video thumbnail (Frame 5693) Video thumbnail (Frame 6907) Video thumbnail (Frame 8305) Video thumbnail (Frame 8716) Video thumbnail (Frame 9805) Video thumbnail (Frame 10949) Video thumbnail (Frame 12763) Video thumbnail (Frame 15395) Video thumbnail (Frame 15858) Video thumbnail (Frame 16580) Video thumbnail (Frame 17240) Video thumbnail (Frame 17656) Video thumbnail (Frame 18709) Video thumbnail (Frame 20550) Video thumbnail (Frame 21482) Video thumbnail (Frame 23387) Video thumbnail (Frame 29397)
Video in TIB AV-Portal: DigitalGlobe and Open Source

Formal Metadata

DigitalGlobe and Open Source
Title of Series
CC Attribution - NonCommercial - ShareAlike 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 and non-commercial purpose as long as the work is attributed to the author in the manner specified by the author or licensor and the work or content is shared also in adapted form only under the conditions of this license.
Release Date
Production Year
Production Place
Seoul, South Korea

Content Metadata

Subject Area
Adding some important and pertinent information to this abstract with respect to the recent tragic events in Nepal. DigitalGlobe, in responding to the devastating earthquake in Nepal, has openly licensed both pre-event and post-event imagery, as well as openly licensed the results of our Tomnod campaign, which has crowdsourced information from nearly 50,000 volunteer contributors to assess damage and displaced people in Nepal. DigitalGlobe is working with first responders, aid relief and NGOs including Kathmandu Living Labs, Humanitarian OpenStreetMap Team, the UN, IFRC and American Red Cross. We are providing important information to disperse relief to the growing number of displaced people. The work we have done has been featured by CNN, CCTV, Mashable, the Atlantic and many more. Please reference links below. This is a poignant example of how Geospatial data, provided in the open can benefit millions of people who need help. Original abstract: DigitalGlobe operates a constellation of high resolution, high accuracy satellites. Imagery from DigitalGlobe can be seen in Mapbox Satellite, CartoDB, Google Maps, HERE Maps, Bing Maps, Apple Maps and is often used for the purposes of contributing, editing and validating for OpenStreetMap. Over the years, DigitalGlobe has provided both imagery and software processing tools with an Open Source license. This includes post-event imagery for Typhoon Haiyan in the Philippines and the Japanese Tsunami. Recently, we open sourced a software toolkit called "Mr Geo" This presentation will give an overview of DigitalGlobe, our geospatial technology and our services we are providing to the Open Source community.
Word Computer animation Open source Vector space Multiplication sign Program slicing Self-organization Open set Product (business)
Word Computer animation Mathematical singularity
Point (geometry) Collaborationism Multiplication Texture mapping Arm Open source Code Software developer Combinational logic 3 (number) Complete metric space Electronic signature Computer animation Software Vector space Term (mathematics) Different (Kate Ryan album) Quicksort Geometry
Algorithm Computer animation Software Open source Vector space Set (mathematics) Text editor
Medical imaging Word Texture mapping Computer animation Open source Vector space Code Right angle Line (geometry)
Word Texture mapping Process (computing) Computer animation Open source Term (mathematics) Cuboid Text editor
Area Medical imaging Image resolution Square number File archiver Point cloud Mass Entire function
Medical imaging Pixel Computer animation Term (mathematics) Different (Kate Ryan album) Image resolution Mass Macro (computer science)
Point (geometry) Building Dependent and independent variables Texture mapping Open source Information Interior (topology) Multiplication sign Online help Water vapor Perturbation theory Event horizon Computer animation Lecture/Conference Internet service provider Electronic visual display Digital Equipment Corporation Computing platform
Dataflow Algorithm Context awareness Texture mapping Software developer Chemical equation Multiplication sign Shared memory Bit Price index Computer Graph coloring Performance appraisal Medical imaging Photographic mosaic Uniform resource locator Computer animation Personal digital assistant Point cloud Quicksort Computing platform
Medical imaging Texture mapping Computer animation Decision theory Interpreter (computing)
Medical imaging Photographic mosaic Building Single-precision floating-point format Interpreter (computing) Endliche Modelltheorie Data structure Computing platform
Medical imaging Texture mapping Population density Computer animation Term (mathematics) Mathematical analysis Boundary value problem Data structure
Type theory Process (computing) Computer animation Traffic reporting Limit of a function Spacetime
Point (geometry) Group action Information Decision theory Online help Bit Water vapor Line (geometry) Medical imaging Casting (performing arts) Different (Kate Ryan album) Order (biology) Quicksort
Point (geometry) Beta function Divisor Open source Multiplication sign Execution unit Open set Revision control Medical imaging Automation Selectivity (electronic) Computing platform Area Algorithm Information Assembly language Software developer Feedback Expert system Computer animation Vector space Linearization Text editor Figurate number Quicksort
at this show slice anyone needs to see the
full the VAD so my name's kevin from from DigitalGlobe among our product team and today I'll try to make up the time which just lost but here to just explain some of things were doing in in open source and open data so jump around and between actually vector in Rasta stuff doing is actually nice that organizes steadily after mark because word we're doing a lot of the the things also today on Amazon and this is me on
Twitter and get out funny get hold of me and so Mr. with food many has anyone heard of this word OK Andrew Mark I didn't know this actually means that the English uh definition of food many that's pretty good on
so yet I can see the full thing but it's a social gathering of featuring folk singing and so actually pretty interesting we decided to name and again have projects and
after it for whatever reason so much at the best name in terms of marketing but people remember it so that she works pretty well so he could get help we have to do this in collaboration with the National Geospatial Agency in the US in GA so the code is completely open source is written by the should developers and when it is is
is completion software that so you can see a whole thing here sorry but essentially and this is a vector based software said where 1 can take multiple vector maps and the self for all actually interpret them and complete them together to produce a 3rd of map which a combination of the 2 and the idea here is to pick the truth or pick the best arm and it's both actually based in geometry geometry based so you know you can address and anything multiple different ways and and often people have 10 different ways of addressing a certain point this software will actually um interpret that and try to pick the best solution are most its automated but and if you wanna run fully automated you can but we we introduced a um sort of a uh manual check even signature going in verify things as well so just show you couple screenshots of this
1 this is all based on the ID editor software
themselves minimize this new you'll notice that the look and feel is a lot like ID I mean this is an example of 2
different sources the blue is what we're calling the source dataset it's OpenStreetMap and the red or a set of vectors somebody has some other traced or and brought in themselves or don't try to complete them together and so the the algorithm will run it will give you a certain checkpoints like this and say well what you think of
this line is this right and you can bring in Rasta images you want to verify it you get to edit the line if you need to move it in a certain way and save your data and at the end of the day or the in the run it will produce
a single vector set a single map so to speak and so on then 1st of all make sure you're all aware of this and word were starting to do all of these projects and and and and make the code available open source and then it goes to show you you know people know digital think of us as a satellite company source company thing he said that's fine but were doing a lot more than just that so matching can give you a few more
examples and many of you are probably familiar Austina data through
OpenStreetMap so if you're in idea editor you're doing your work you you pull in an image and In IT editor for jobs until he simple and being imagery Europe Mapquest imagery map box I'm most that she comes from DigitalGlobe so we licenser imagery of to being they make it available to them in cooperation with us to OpenStreetMap so if you use OpenStreetMap you probably used a sugar imagery and that's great but were were very happy with all the words that that is taking place the OpenStreetMap and more excited about the future but I want to give 1 example that was really powerful and this year in terms of our our work in the open source community and then that wasn't that the earthquake in a
Paul on this big square that represents about a half-million square kilometer area over the whole I'm seeking napoles in here in the middle this is all the images were collected after the earthquake so that was a massive amount of data I unlike Landsat went and made uh when you acquire things a 30 centimeter or 50 centimeter resolution the data is huge or the data are huge from being precise remark on so we're talking terabytes of data from or entire archive is close to hundred petabytes of data are biggest problem right now is we have too much data so that's where Amazon S 3 actually comes in and helps us out a lot so after the Paul and we started collecting imagery the day after a while
and so the earthquake happened on Saturday this is the image from Sunday and it was it was cloudy and rainy but for whatever reason the whole opened up in the clouds and we were able to pointer satellite through that hole in Katmandu actually right right here on what's interesting is is
we design satellites were not a my Krycek company were macro set company enough as a term for satellite big because that massive telescopes and it's really just photography 101 need a big focal length to and get a high resolution pixel but also a satellites can maneuver so what we're doing while the weather was bad and Paul are satellites would pass over in in some the images we actually be over the Indian Ocean or Bangladesh pointed back looking at a Paul and that actually worked beautifully and in fact it was it is very satisfying all that work we but years and years and building the satellites actually worked on it actually save lives
estimated difference but the way we made a difference
is by releasing it open source and involving a big crowd of people the so hot poets and there's about 6 thousand volunteers we launched a we have a crowdsourcing platform called Tom nod and we want did we had 60 thousand volunteers from time not the way I described it's like it's a crowdsourcing platform for anyone my 7 year old daughter can do it my nine-year-old grandmother can do it they can do OpenStreetMap editing but the candy Tom not is very binary that DEC attend yes or no so we lost a big campaign and had 60 thousand volunteers millions of data points we release that is open source as well and
so smart people smarter than I this is a map made by Miller Bill Morris yet she took the OpenStreetMap data the town I data them together and it helped inform the work we're doing after after the earthquake and it was really interesting to me in this in this scenario is we were looking for toppled buildings everyone knew was toppled buildings were we're looking for displaced people and the internally displaced people and so we're looking for these camps that were building up after the crises to make sure people and for display said food water and medicine so wasn't really the earthquake it was the response of the earthquake the I the in the polls in the past that the important thing is it were building and infrastructure in a community for what happens in the future and this can be another event somewhere in the world um where people need help and would you know drives me makes me you know happy going to work every day is that we can actually from the inner city in Colorado we can actually provide data and information around the world to people in need of most of them and we do that through open source and recently launched a
developer portal and you can go to develop adopted a little dot com indices see a couple API is there and what is the Maps API so 1 March of the slippy map that's exactly what this is it's built on the map OX CPI but it allows time you access a global 30 centimeters color balance mosaic of the world in which again there is a hard thing to do but it's all based on Amazon S 3 and you can sign up for an evaluation if you wanted to actually subscribe we just charge by this you can buy buy transaction and so we have a global map of the world available there and we also have something and down here it's a little bit cut off of ecology especially data so is the problem of having too much data and we don't want a ship if someone were to say I need to solve you know that some sort of global problem we don't ship them hundred petabytes of data we say go to the cloud hidden API and bring your algorithm I think Marx said don't copy the data that bring your and bring your work flow into the cloud to exactly we're doing so through g especially data you can see every image we've ever collected that goes back to 1999 from all around the world you have access to it all in the cloud there's no such thing as delivering data through this platform you can view it you can analyze it we have a bunch of algorithms you can run on and off your bring your own algorithms you can't slot so I'm really interesting use cases there are being developed at that so I just wanna share fewer and then from market and a home and doing on time but I'll share fewer men of enough for questions you about 10 minutes OK so
this the principles behind you spit geospatial Big Data is everything happens at a certain location at a certain time in this budget context around that that we need to understand that and this can be passed or or future and and and
so like use a beautiful at the image of the river and but what's really important is the interpretation of that and and actually building out of land
use maps you can see the legend here of what actually in that image and I see imagery almost as a novelty almost like the yeah it's great to look at what's really important are the things within an image that's said that's what makes us allows us to make better decisions so and a just a few
more but this is actually Tokyo Japan and so an image with the kind of stuff in it we can actually start running interpretation models and find an individual buildings so we're we're able to take a uh uh and imagery mosaic of all Japan run a model on it and find every single building of a single structure within hours using this platform so I'm again here the
image at this is somewhere in Africa back to this is in Somalia and we're the when you
look at this image again it's it's interesting but was telling me if you run some analysis on it we can actually find a population
density terms of this is a dense population density map of that image the so that such what's important if you're a humanitarian agency or government and even it's amazing were spoiled in places like that Korea and the United States we have a robust government infrastructure with a census most the world doesn't have a census and so geospatial data is a great way of of that creating a census so once you know population and city and these are the village boundaries and these are the actual uh homes and and structures within that and then scale it up to the Horn of Africa
and so we can find everywhere where people live and quantify it can and there's many reasons why would I'm recently
and we are able to find a uh a slave ship in improper New Guinea and this made the news a couple weeks ago uh we're working with an investigative reporter and we found the slave ship from space and it turns out that these 2 this is a commercial fishing boat in these 2 boats are here slave boats what happens is they have if it's remarkable that that this happens but there are hundreds of thousands of people they're enslaved today what happens is that these are Burmese men were lured but with the with the hopes of having a good job and then approach could work on a slave ship work 22 hour days have barely get enough food to eat to survive but they they were stuck out at sea for a couple years we found them and so 8 men were actually liberated because of this and this is amazing like these are the types of things that were able to do so I'm using the that's cool satellites and then geospatial analytics and and the crowd of people to help us out the last 1 this
one's and but if you turn on the news you've seen the refugee crisis so this is a really hard 1 to solve were actually still trying to figure out how to solve it and in 1 of the reasons so I can this conference was not to show you how smart and cool we are but is really like enlist help and and thoughts some here more to learn than I am to represent anything but right now you have hundreds of thousands of refugees from Syria in Afghanistan I'm trying to find safe haven so this is the border that in Hungary are from a year ago and we were able to look at it and just
from a few days ago and they've actually um they've actually could created a razor wire fence you can actually see the difference in the 2 them and this is the border so the creator razor wire fence to prevent migrants so which is a whole nother story I won't get into but the point here is were able to fish to show that ensure what's going on and and actually in this image here is a border and here's a truck that and I think it's like at border than you know some sort of order each gone up in in the water and you can actually see a group of people and those the individual people there were trying to get through this fence and now that Ch I'm not sure what to do with this information but were collecting it and we will make available and and and hopefully all of us can make sense of it and that I think that's the the biggest point here is we have tons of data we have made data we have Landsat we have different little data we have earth cast is coming line on how we get together and make sense of it all and make make good decisions and so that's hopefully a little bit of like inspiration to and my talk but that's all I had today and so thank
you and I'm happy to take questions in last whatever minutes we have and at
this point yeah you yes the the the and validated you that Yuki's faster data yet d dt that automated way or is a manual verification of so we have a way of doing it in an automated way and it's essentially uh an algorithm that looks for linear paths or roads and then can can pull vector to align with that road it's not always the best way of doing it and the with food many what we see most are editors in who they the the experts of that in a particular area actually doing manual editing themselves and in this can like that human that curation of the dataset which is probably always can be better than automated so is a really good question how we can actually do both and and going back that geospatial Big Data platform actually put both those algorithms on the platform seeking pull-down image and the algorithm we have it's called road tractor select tracker in in this sense extract every road within an image and and then you can you can edit or whatever yeah but but again great great question and that's like to me that's the exciting thing of how we pull features and in information out of an image the so there's no more questions I'm around all week and please spare developer portal and you can ask us questions ask me questions and but you know please provide feedback in the end you know that's that's really my main reasons I came here so thank you for the question you know that when the size of 3 point 0 you and that there's been need some of this panel recently by the European Space Agency The Novel not next year yeah have read something that they were gonna open source of the data in the five-day refresh his that like a cantonal xi the sensor 1 yes and you get a thing with beta form here if that based comes open of divide unit in fact yet a good question and so I think the sentinel would be comparable to our Landsat him someone can fact-check me on that but I think the Sentinels can the European version of the Landsat program by the the answer is it within our platform we actually pulled in all of the Landsat dataset that we pulled in time a search him which is the uh it's from NASA's shuttle radar topography missions basically and elevation knowledge of the entire world and so you the answers when data becomes open source we do pull or platform so you're not only looking at imagery from our satellites but from other satellites as well and so the entries let's figure out what satellite that is and and OR apart from the data agnostic so it doesn't matter if it's you know they come from UAV or um assemble satellite we can we can deal with it but but actually if all philosophy is it's better to aggregators much sources as possible and because no 1 sort source could ever be the truth so we pull multiple sources and that tends to work better for our users OK thanks thank you so much the aaai 1148 of 2 minutes to make that I don't think there's a mother talking here I see of 2 minutes go making accession thank you for your time today thanks for coming the kind