OpenAerialMap V2 Design and Development
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
Title |
| |
Title of Series | ||
Number of Parts | 351 | |
Author | ||
License | CC Attribution 3.0 Unported: 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. | |
Identifiers | 10.5446/69123 (DOI) | |
Publisher | ||
Release Date | ||
Language | ||
Production Year | 2022 |
Content Metadata
Subject Area | ||
Genre | ||
Abstract |
| |
Keywords |
FOSS4G Firenze 2022121 / 351
1
7
13
22
25
31
33
36
39
41
43
44
46
52
53
55
58
59
60
76
80
93
98
104
108
127
128
133
135
141
142
143
150
151
168
173
176
178
190
196
200
201
202
204
211
219
225
226
236
242
251
258
263
270
284
285
292
00:00
Perturbation theoryTexture mappingSelf-organizationOpen setRevision controlStreaming mediaMobile appRight angleMultiplication signComputer animation
00:55
Derivation (linguistics)Open setCollaborationismDependent and independent variablesSet (mathematics)Perturbation theoryStrategy gamePresentation of a groupSelf-organizationTexture mappingStreaming mediaBitWebsiteSoftwareOpen setWeb 2.0Web pageDependent and independent variablesInformationSlide ruleMereologyRemote administrationActive contour modelMultiplication signMobile appDirection (geometry)SatelliteComputer animation
04:40
SatelliteOptical character recognitionExplosionMedical imagingTexture mappingWeb browserInterface (computing)Physical systemOpen setAreaSoftwareCASE <Informatik>TesselationComputer animation
05:20
Product (business)ExplosionSatelliteBit rateEvent horizonLoginStreaming mediaCASE <Informatik>Medical imagingDependent and independent variablesSet (mathematics)Computer programOpen setMetadataSatelliteTexture mappingCartesian coordinate systemAreaComputer animation
05:54
TensorData typeComputer-generated imageryComputing platformSatelliteOpen setMetadataBlogUniform resource locatorMobile appMedical imagingMetadataComputer configurationOpen setTexture mappingSet (mathematics)Component-based software engineeringInformationComputer animation
06:23
EmailAddress spaceGaussian processData typeMixed realityCycle (graph theory)ForestWeb browserSystems engineeringCoordinate systemLambda calculusGateway (telecommunications)Open setText editorCASE <Informatik>Streaming mediaMultiplication signService (economics)Mobile appTexture mappingInstance (computer science)TesselationRevision controlLambda calculusDataflowComputer animation
07:14
Cloud computingWeb browserMathematical optimizationTesselationServer (computing)Data storage devicePhysical systemCloud computingData compressionService (economics)Multiplication signOpen setFile formatRaw image formatMobile appTexture mappingComputer animation
07:58
Internet service providerVertex (graph theory)Stochastic processInterface (computing)Computer-generated imageryWeb 2.0Open setTexture mappingMedical imagingPresentation of a groupCore dumpPlug-in (computing)SoftwareComputer animation
08:41
Exception handlingComputer networkOpen setCondition numberLatent heatSynchronizationComputer-generated imageryDatabaseMachine visionStrategy gameSoftwareSoftware testingArchitectureProduct (business)Point (geometry)Statement (computer science)Texture mappingPrototypeSet (mathematics)MetadataProjective planeTexture mappingVisualization (computer graphics)PrototypeProfil (magazine)Mobile appPhase transitionSoftware testingMultiplication signMereologyFeasibility studyStack (abstract data type)Surjective functionPlastikkarteComputer configurationMathematicsPoint (geometry)Ocean currentFlow separationRevision controlCodeImplementationComputer architectureOpen setPublic domainCartesian coordinate systemInformationSubject indexingDataflowReal numberSinc functionComputer animation
13:04
SoftwareOpen setSatelliteSelf-organizationHuman migrationPerturbation theoryShape (magazine)Meta elementComputer hardwareComputer-generated imageryVolumeTask (computing)Texture mappingDecision theoryMobile appElectric currentPoint (geometry)Medical imagingType theoryShared memoryData managementProcess (computing)Phase transitionStochastic processCartesian coordinate systemDataflowPoint (geometry)Projective planeGroup actionVirtual machineInformationDifferent (Kate Ryan album)Error messageSoftwarePlastikkarteReal numberMappingOcean currentImage resolutionLink (knot theory)Physical systemCodeComputer animationDiagram
16:38
Dean numberConditional-access moduleSatelliteVotingTexture mappingGEDCOMComputer-generated imageryImage resolutionStatisticsBinary fileData typeComputer configurationAreaFingerprintArc (geometry)Square numberMilitary operationElectronic visual displayUniform resource locatorText editorMedical imagingDecision theoryInformationComputer configurationFiltrationZoom lensWeb pageAreaSystem administratorPoint (geometry)Texture mappingCellular automatonContext awarenessFreewareTouchscreenElectronic mailing listMultiplication signNumberBitBoundary value problemGeometryPerspective (visual)Gene clusterComputer animation
19:25
Personal digital assistantOpen setComputer networkCondition numberException handlingInternet service providerMetadataOpen sourceData typeData modelGoogolComputing platformSatelliteImage processingAerodynamicsMathematicsOpen setMultiplication signTexture mappingQuicksortData managementComputer architectureGeneric programmingAreaMedical imagingField (computer science)Active contour modelUser profilePhase transitionOpen sourceComputer configurationReal numberDynamical systemStack (abstract data type)ImplementationComputer animation
22:25
Continuous functionPerturbation theoryStrategy gameComputer configurationProduct (business)Reduction of orderUsabilityMetric systemProjective planeLine (geometry)Beta functionImplementationTexture mappingMultiplication signData managementComputer animation
22:49
Link (knot theory)WebsiteBlogAreaService (economics)Self-organizationTexture mappingTouch typingActive contour modelFeedbackOpen setProduct (business)Point (geometry)Musical ensembleComputer animation
Transcript: English(auto-generated)
00:01
Thank you, everyone. We are ready to start. Right on time. My name is Cristiano. I'm with DK from humanitarian open stream app team. You might have heard of the organization and we have also Katharina from Condor. Today we're going to talk about open air map and about open air map B2, or the new open air map, which is in the making.
00:25
So we're going to start with a quick introduction about what is open air map, why we made open air map, quick history, which has become quite long already, and now we are doing an effort to bring open air map to a new version to provide more features, modernize the infrastructure.
00:48
So DK will talk about the infrastructure and Katharina will talk about the design and the design approach that we took in this. So, first to start, we heard from Piero about
01:03
open software and now we're talking about open data and why open imaging particularly matters. We are an organization that heavily works with open stream app data and so one of the things that I listed here is that it allows open directive data to be created.
01:24
So we do a lot of data creation in open stream app with humanitarian open stream app team and we need the imagery from which we derive the data to be openly licensed. So it's really important for us that the imagery that we use has that, is open and
01:42
freely accessible. There are other reasons why open matters for imagery. One thing that I think is important is it also provides access to transparent information that can be used for AI ML work. It's there, anyone can verify because the imagery is available
02:04
to anyone. So, recapping quickly from 2006, how open air map works. Open air map was started. 2006 was the first years when drones or remote control airplanes
02:23
were started to be used with digital cameras to capture aerial imagery. So the beginning of the drone technology for consumers. All the way to 2010, humanitarian open stream app team started around that time responding to the Haiti earthquake and there was a lot
02:45
of imagery made available by satellite imaging companies to respond to the earthquake and we needed a place to put that imagery to trace and create data in open stream app for the responders. The idea further developed until 2014 when hot humanitarian open stream
03:04
app team submitted a proposal, won a grant and in 2015 were able to start the design and the development of the first open air map. So Kate Chapman, who was here earlier in the conference, presented a soul in 2015 of Phosphor-G, the initial idea of open
03:24
air map. Since then, many other organisations and people that are here today have been contributed and I don't have a slide with all the contributors like Pierre attempted
03:40
for drone DB, but they are on the website for open air map and the about page. Since then there has been some little improvements but pretty much what we have in open air
04:06
map is that we are able to build a software. So it's a little bit outdated. In 2021 we had some funding or I was able to secure some funding to work on redesign of open
04:20
air map and so that's where we are today with contour. We started working on the design part and the development is going to start soon with a target of the early next year to have it functional and so DK will also talk about that. Quickly, some of you might
04:42
have heard or even used open air map, but the idea was to make a system or a distributed network of images or image nodes and create an interface to be able to browse and create access to data and upload imagery to contribute imagery to open air map. In this case we
05:07
are looking at what the browser looks like today. It's a map with tiles that show based on the shading of the blue where the images are. Once you click on the images
05:21
then you can access the image and download it or stream it, we will see it. There are most of the images in open air map are from drone mappers, so you will find a lot of those, but we also have satellite imagery that has been released mostly for
05:40
humanitarian applications or response. In this case we have maxer imagery that was released through the open data programme as a compatible open licence. It was uploaded to open air map. There is a set of basic metadata that needs to be entered when you upload the image into open air map and so there is this component which is the uploader
06:05
where you do that and when you zoom into one of the images in open air map you have the option to download it to directly access the image through TMS, WMTS and you see basic metadata that tells you information about the image.
06:24
In this case we are seeing the same image that we saw before inside the ID editor, the open stream app editor coming in through the time map service and the same image in QGIS through the WMTS service. The idea was to build a simple and highly
06:45
available infrastructure. Right now open air map, the current version is hosted on AWS and this is the work flow on getting the image up into an instance that processes
07:00
the image, it puts into an highly available bucket, an S3 bucket and then through lambda service, serverless service we provide those tiles through the WMTS and TMS services. What happens when you upload the image into the upload? The raw image gets uploaded to
07:22
that server, gets compressed and optimised, you have heard in other talks about COG and saved on the cloud storage. Open air map thanks to some smart developers that are here at the conference, it was one of the first systems that implemented COG and I don't
07:41
know when it was, like 2017, around 18 and it's still working exactly like it was at the time and serving tiles off of the format after so many years. And we heard in the last two presentations about ODM, open drone map and web ODM, one
08:08
very reliable thing that we have after many years and many things that have broken because we haven't really maintained the core software, but one thing that really works well is
08:24
the connection with web ODM. You process the images in web ODM and automatically it gets uploaded through this plugin to open air map, so when you have your drone images you process them, create a mosaic and then it gets automatically uploaded into open air
08:40
map. There are some things that we are hoping also to address as part of the development and design that we are doing. Two key things, one is about the imagery licensing, there are different efforts to try to have a set of licences for imagery or for EOData, but many times we are trying to make sure that
09:06
we are still using Creative Commons, ODBL or public domain, but I think it's important that we start thinking about a licence specifically for EOData or even better for imagery.
09:21
And the same for metadata, there are things that are being developed now as part of stack, there are efforts within ISO GC to have those profiles, but something simple that we can easily fill out, but someone that doesn't have that background in metadata, if that's
09:43
even a background, it's something that we need so that we can easily have a drone pilot upload information that we need to be able to index and search that imagery. So why the redesign? I think I mentioned it, it's been many years since we first developed
10:05
the first version. I think we still have many more people happy that they are able to upload the imagery onto OpenAI map and use it for their different applications also beyond humanitarian applications than those that are complaining that things are broken
10:22
because yes, we haven't been able to really fix some of the code since then, but it's working fine exactly on that technology from the time. So the goal is really to increase accessibility, modernise the stack and so we'll see now
10:46
with the approach and the design and then the proposed implementation options how we're going to do this. So I'll leave it to Katerina for the design part. Thank you Christiana, hope you can hear me. So I will tell you about the redesign approach
11:11
that we applied to this project and about main outcomes and some insights. So we applied design syncing approach for this project because we didn't know at the
11:23
beginning what's the main problem for users and we needed to extract this information and find the more paying points on the user workflow. And we have had four phases, you
11:40
can see it here. The first phase, discovery, we identified the main stakeholders who can be benefited by open aerial map redesign, then we interviewed them to gather main user flows and create user personas to easy work with user needs and common goals.
12:02
Then the second phase we created some design proposal, it was a lot of ideas how to improve open aerial map, but we prioritised these ideas and for most viable of them created interactive prototype. And the testing phase we showed our prototype to real users
12:30
and checked if they really understand what we, if it's really better for them. And the last phase we checked the feasibility of the proposed solution because of course
12:46
current architecture doesn't support some features and it was designed several architectural approaches, how better to implement these changes. So let's go to some more cute visuals
13:04
maybe. And the first phase as I said we gathered information about main stakeholders, there is two main purpose for this. First one is to invite workshops and introduce just relevant people who can give really some actual information, real pain points.
13:26
And the second purpose is not to miss anyone, to cover as more user groups as possible. Actually we covered almost all of them, just maybe government was not covered properly
13:40
but we tried. And you can see some names of the company we reached during this process. And then when we had some talk with the main stakeholders we realised that there are some main general user flows and created based on this information such as user personas,
14:06
user personas are fictional characters so they present some let's say maybe type of user
14:20
and for each user we gathered current flow, needs and pain points. Here you will see just three of them because we concentrated only on these three personas so it can be of course more but during this project we concentrated on these.
14:40
First person is mapping project manager, is person who manage coordinates, work around mapping and search for best images, maybe recent one and usually wants to share this
15:03
information using some link code to use this information for different external applications and tasking manager for example and main need of course is access as many images as possible and find some relevant.
15:22
Second person is imagery contributor, is person who brings his data to the system and this persona wants to just quickly upload his information without any maybe long processing
15:42
and some errors because it can lead us to this persona can leave for the application because it's not very beneficial from the application, it's only want to upload, share
16:06
some information. And the last persona is GIS analyst, of course it's quite similar as first one, but usually they have more exact requirements for data, maybe some resolution or some data
16:29
and this persona downloads images to local machine and then processes it using different software. So for each persona we gathered, we call it pain points, problems that these personas
16:48
have and we just will see some current situation and some proposed solutions for this persona for Ux perspective. So now you can see main screen where people see some coverage by images and it's a grid
17:04
cell and usually it's not so transparent how many images in your area of interest there is, so you zoom in, zoom in to your maybe city and yeah it looked pretty good before but when you zoom you can see just several images, so we proposed some
17:25
clusterisation, you can see number of images without any hovering and also you should see left panel connected to the map because the current solution you cannot do it.
17:43
Second scenario is filtering, now filtering options is a bit poor, you can see this here, so last week, last month, last year, it's not these filters that users need and there are some proposals about extending the filter list and also introducing timeline as a way
18:07
to filter images by dates because it's convenient and easy to use. Next user scenario is searching by area of interest and now you can also search by images
18:22
within the grid cell and you cannot see any images on the map, just on the left side. So we proposed to filter by area of interest using some search by administrative boundary
18:40
or even free hand geometry, of course it's a good idea to show all images at the same time and highlight the area of interest to show it visually clear. Actually image page is not so painful, just for some decisions there are not a lot
19:01
of information and you cannot see any context, so you can see just one image at one time. So we can add more data, all data that we can get from the uploading and show if there
19:20
are any images near this one, it gives more context and all previous pain points was related to people who consume images but person who uploads images also faced with some issues, usually this is not clear how to fill in some fields and sometimes
19:42
uploading is crashing and it's not so clear why. It's usually some classical approach to add some more hints, maybe rename some fields and then we decided that some fields can be prefilled and even removed and we added
20:04
some suggestions about how we could check images before uploading, just not to crash. So because of all the previous improvements that we introduced of course some changes
20:24
in architecture is needed, there is some generic architecture of open area map and I'll give a floor to Dike, he will tell you in details about architecture changes. Hi, so we're going to real quick talk about architecture, I don't have a lot of
20:43
time but as you can see here this is broadly not too unchanged from how OAM is currently running but to really sort of make sure we're implementing some important technologies that will support the design including user profiles, actually being able to manage user
21:04
profiles and give these images a connection to a person as well as of course implementing the stack API for interoperability, for the ability to share across outside of even
21:20
open area map. We have a dynamic Tyler, it's about 7 years old and there's been a lot of changes to how dynamic Tyler is built, so we'll just look at some finer details on how this might be implemented, some guidance from the team. You probably have seen many of these logos,
21:42
there are so many open source solutions and open area map is a stack of tools that relies on many currently and will continue to rely on this community. And then of course we do run on AWS so there are some ways of running all of this on AWS as well. And let's see, last one, yeah with Contour as well, they have lots of tools
22:07
that they have built themselves and so we are looking at, the guidance that we were given through all of these is sort of just to show that there's a lot of options, we have a lot of work to do to research and understand the best way to implement open
22:22
area map in its next phase. Just a quick roadmap, what's our time? We're good? Okay. This is a very loose roadmap that we developed just to honestly for a lot of things just as a project management, but in the second line we're looking at Q1 2023 for some first beta implementation
22:46
of the technology. And lastly just some ways to get in touch, if you are a user of open area map please get in touch with us, tell us your stories, tell us your successes, tell us your pain points, we are still open to a lot of feedback and then also please get in touch
23:03
with Contour as well. It's been I think a huge success story for us as an organisation of being able to work with a partner to do something like design, a lot of technology maybe sometimes comes into play from honestly happenstance and so when we're given the opportunity to
23:20
work with people and do something like this, I think we're able to create really great products that provide a great service. Thank you.