VectorTile for SeaWind
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 | 295 | |
| Author | ||
| Contributors | ||
| License | 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. | |
| Identifiers | 10.5446/43386 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
| |
| Keywords |
00:00
Vector spaceProcess (computing)Demo (music)Integrated development environmentMiniDiscSurfaceProcess (computing)Service (economics)Presentation of a groupVector spaceType theoryInformationObject (grammar)Ferry CorstenComputer animation
00:32
Process (computing)Vector spaceDemo (music)MiniDiscIntegrated development environmentDimensional analysisTouchscreenDemo (music)Presentation of a groupDirection (geometry)Demon
01:02
Vector graphicsInterpolationTouchscreenDaylight saving timeBoss CorporationVector spaceMappingComputer animation
01:40
InterpolationVector spaceBoss CorporationMedical imagingOffice suiteOperator (mathematics)BitBeta functionSurfaceVector spaceType theoryInterpolationResultantSymbol table
02:24
Vector spaceAreaVector spaceProcess (computing)MappingElement (mathematics)TessellationComputer animation
02:47
Vector spaceTessellationFile formatBound stateAxonometric projectionNumbering schemeGoogolCommunications protocolBuffer solutionFile formatFormal languageData modelComputer animationJSON
03:07
File formatVector spaceBound stateAxonometric projectionMaizeGoogolNumbering schemeField (computer science)Direction (geometry)Regular graphImage resolutionProjective planeTesselationBound stateVector spaceType theoryNumbering schemeAreaLevel (video gaming)Latent heatPhase transitionInformationRow (database)File formatBijectionExtension (kinesiology)Combinational logicDatabaseProduct (business)Game theorySpacetimeGeometryMultiplication signSet (mathematics)Coefficient of determinationWeb 2.0Source code
04:09
Field (computer science)Image resolutionRegular graphDirected setDirection (geometry)Regular graphDirection (geometry)Image resolutionTask (computing)Lecture/Conference
04:24
Image resolution8 (number)Image resolutionCAN busSpacetime
04:49
Field (computer science)Image resolutionJava appletSpring (hydrology)Integrated development environmentVector spaceAxonometric projectionConformal mapSatelliteASCIICountingTurbo-CodeProjective planeMultiplication signPiProcess (computing)BuildingElectronic visual displayInternetworkingForestLibrary (computing)CalculationIntegrated development environmentVector spaceTessellationSpring (hydrology)NumberWeb pageTimestampElectronic mailing listComputer fileType theoryOpen setComputer animation
06:30
Reading (process)Coordinate systemBound statePoint (geometry)Mean value theoremWritingComputer fileCalculationSinguläres IntegralElectronic visual displayPoint (geometry)MappingCoordinate systemElectronic visual displayFile formatCodierung <Programmierung>Form (programming)TesselationBoundary value problemBoss CorporationMereologySystem callRoundness (object)JSON
06:58
Zoom lensBuffer solutionLevel (video gaming)Service (economics)
07:14
Point (geometry)Web browserScale (map)Image resolutionPairwise comparisonDistanceLevel (video gaming)ResultantZoom lensTable (information)LengthImage resolutionExecution unitComputer animationProgram flowchart
07:39
DistanceReal numberImage resolutionPairwise comparisonMappingTable (information)TesselationExecution unitDistanceImage resolutionLengthMultiplication signProgram flowchart
08:14
Level (video gaming)
08:29
Point (geometry)Boundary value problemMean value theoremPolygonAreaSample (statistics)CodeVector spaceAxonometric projectionData bufferCalculationSource codeVector spaceCodeLevel (video gaming)TesselationTessellationBuffer solutionUtility softwareCoordinate systemData conversionLibrary (computing)Point (geometry)MeasurementWebsiteService (economics)DivisorComputer animation
09:18
Host Identity ProtocolCodeData bufferCalculationAxonometric projectionPixelMetreImage resolutionFunction (mathematics)EstimationCodeTessellationLevel (video gaming)MathematicsTesselationFunctional (mathematics)Projective planeProduct (business)Computer animation
09:39
MetrePixelImage resolutionFunction (mathematics)Data bufferCoefficient of determinationMetreBoundary value problemCoordinate systemImage resolutionPixelTesselationBuffer solution
10:08
Loop (music)Point (geometry)Price indexDensity of statesForestSemiconductor memoryPoint (geometry)Set (mathematics)Channel capacityHierarchyMultiplication signBasis <Mathematik>Data structureQuicksortIterationTesselationBuffer solutionGreatest elementTessellationComputer animation
10:45
Loop (music)Price indexPoint (geometry)Electronic visual displayVector spaceType theorySubject indexingPoint (geometry)GeometryTesselationOpen setComputer animation
11:12
Electronic visual displayVector spaceSource codeFile formatImage resolutionDemo (music)Social classDefault (computer science)TesselationParameter (computer programming)PixelExtension (kinesiology)Mean value theoremKey (cryptography)
11:42
Lecture/ConferenceComputer animation
12:22
Goodness of fitComputer animationLecture/Conference
13:44
EmailOffice suiteCausalityPhysical systemProjective planeSource codeWeb pageFrequencyType theoryDesign by contractStandard deviationWebsitePoint (geometry)Multiplication signVisualization (computer graphics)Presentation of a groupClient (computing)SmoothingBoss CorporationProcess (computing)Sound effectMathematicsReal-time operating systemTotal S.A.Instance (computer science)SatelliteInternetworkingPosition operatorDivisorBitLecture/ConferenceMeeting/Interview
Transcript: English(auto-generated)
00:07
Hello, I'm Yona Zhang in GAIA 3D. I introduced the process of giving service using vector type for displaying sea surface wind data.
00:20
Okay, the presentation is as follows. First, I will tell why I used vector type and describe what is the vector type. Next, I will introduce the information and process necessary to actually build a service. Finally, I will show the live demo.
00:46
Wind data is very complicated because there are wind direction and wind speed. Wind direction is three dimensional data, so adding speed, there are four dimensions to represent, but the screen is displayed in two dimensions.
01:08
So, four dimensional data cannot be represented on the screen by itself, so must be divided into two dimensions. Generally, like the left picture, one vector is represented.
01:24
However, like the right picture, you can represent most of the data at once by coloring the vector of the wind maps. This is the first method that I used.
01:47
There is a lot of data that needs to be displayed, so I've chosen the tiling method using bounce surface interpolation, but the result of interpolation covered over land and rotated wind symbol seems to be a question.
02:08
To solve this problem, I used the vector type method. As you know, I briefly introduced vector type.
02:23
Vector type is to think of as dividing and storing vector data into types. There is no rest of the images, only fully generated vector data, which I presented on the request area. Those maps' elements we see over pearls' current tiles are cut with a certain offset,
02:46
which is essential when connecting tiles together. Embedded formats use Google protocol buffers as an encoding format.
03:01
Protocol buffer doesn't consider kind of languages for reference by processing serializing structured data. The vector type represents data based on squared extent within a projection. Vector type should not contain information about its bounds and projection.
03:27
These formats assume that the decoder knows the bounds and projection of the vector type before decoding it. Web recorder is projection of reference, and the Google tile scheme is the tile-extent convention of reference.
03:51
They provide one-to-one relationship between specific geographical areas at specific level of detail and phase like this.
04:04
Now, I pray about the data that I used first ask B. The resolution is traveled 5 km or regular data with latitude, longitude, date, time, speed, direction.
04:29
Detail specs. Second, wind set. This resolution is like this.
04:47
Details. Final, last, scast set. Specs are like this.
05:00
This is the developed environment. ZK 1.8, Spring 4, Maven, Jirutus, Mapbox, vector type Java, open layers. This page has a list of vector type project.
05:22
I was buying the library in Java on this PC. It's good to be able to select library your project environment. The requirement as shown is Pi needs to hold the timestamp and the data projected into the custom project
05:50
and files are merged into the tiles for 12 hours, regardless of processing time of the files.
06:03
Let's calculate. As you can see, it's huge number because you have to merge data accumulated in 12 hours. So what happens if you render this data directly? It's too hard to see.
06:33
The overall values is as follows. Coordinate transformation is performed by reading the coordinate, calculate the bound of the tile
06:46
and find the point contained in the tile, encode the tile in NVT format style in the form of wind maps and display it. I'll explain how to calculate one of the buffer and skip data based on zoom level.
07:14
The first problem is rendering problem at a higher zoom level. To solve this, I selected the result skip data as zoom level.
07:30
The way to skip is like this. Calculate the resolution length of displayed picture and get the interval unit with this value.
07:43
For example, I'll explain it with a table. I calculate interval unit by dividing the real distance among the wind maps into the resolution of the data. I get the length of the tile by using this interval so I could decrease the tile size more than 10 times.
08:16
Solving one problem leads to another problem. The level data has been skipped.
08:23
So I save the level data by dividing from the wind data. If the landed feature is overlapped with only several tiles, the dead feature is cut. As mentioned at the vector tiles back part, I need the data of a feature like this connect its tiles.
08:49
To solve this, I need to give the buffer around the tile. The following is example code provided by the library.
09:05
The example buffer the tile size multiplied by 0.1. I found source code for utility of conversion between tiles and coordinate in map tiles size.
09:24
I reuse the code with global map tiles.py code. The changes are TMS origin and projection. This is the tile grid functions.
09:43
Calculate the resolution of target zone level. Calculate the coordinate is the meter xyz. Calculate meter coordinate from the pixel xyz. Calculate boundary of the tile xyz in meters.
10:01
Calculate the boundary of buffer tile xyz in meters. At the first trial, I iterated through the set of the tiles, but it took too long memory capacity and execution time.
10:25
So I changed the set of the tiles to set of the points. At the second trial, I tried to make the higher rack structure from the bottom of the top, but applying data intervals is very difficult, so I mixed up both of the methods.
10:45
On the basis of points, iterate points and check each of the tiles whether the points hit the buffer of the tile. If it matches, then save the tile index and geometry.
11:03
I used open layer spectra tile layer. Note that the tile pixel ratio value is the extent divided by tile size,
11:21
which is set by default in mvt layer parameter class and can be changed. So let's live demo. Start live demo.
13:48
Thank you for listening to my presentation. If first of all, please email questions.
14:06
Thank you so much. Is there any webpage where we can see this working? She is my colleague.
14:23
She is my colleague. I would like to answer to your question on behalf of her. I don't think so because our client is Korea Meteorological Agency and they have a Korea Meteorological Satellite Center and this is only for the intranet users, so you can't access the site,
14:44
but other officers from Korea Meteorological Agency can access this site and enjoy that. Please, questions.
15:06
Is it frustrating that the source that you're using only has incomplete data for the wind speeds? Do you understand?
15:22
So the data you have seems to have strips. It's not complete. So is that frustrating? Does that cause a problem to be usable? She's working on that. Yeah, there are lots of problems regarding the data and she's working on that.
15:43
The contract is not finished. Any more questions? Yes. Nina, I'm scared. We're talking about wind speeds and you were talking about air balloons. I'm worried about what can happen here.
16:01
Yeah, I might put these air balloons on this data. So what is the update frequency of the points? Do they real-time update? Does it change the static? Oh, this data is three minutes updated.
16:29
So your points change every two minutes? Cool. We still have some time for more questions. If that's fine for everybody else.
16:49
So because I've done work in the aviation industry, I'm familiar with the wind barbs method. Have you looked at other ways of depicting wind? So for instance, in television weather forecasts, there is often a very smooth, animated way of showing wind.
17:09
And it's a very famous website now, but if anyone hasn't seen it, there's a website called earth.nullschool.net, which comes up with a non-GIS visualization of wind, but it's mesmerizing, it's really beautiful.
17:25
So I don't know if your project could consider that type of visualization. Yeah, thank you. I consider it.
17:40
I'd like to ask you a question. The wind barbs method is very similar to the wind barbs method. The point-to-point method is very similar to the animation method. The main use of this system is forecasters in the Korean meteorological agency.
18:00
So they want a professional standard in the system. So they do not want the kind of animated vectors or other things. That's very cool, but they are professional guys, so they want to analyze point-by-point and region-by-region. That's why they just prefer these kinds of things.
18:22
Okay, so any more questions? If there are no more questions, I would like to thank you for coming to Phosphogee.
18:43
This is the first presentation at Phosphogee. You did wonderful. Thank you very much.