Sign in to follow this  
rhumbaflappy

LOD11 span in metres - Urgent!

Recommended Posts

Hi everybody,Does any1 know exactly what the width and height of an LOD11 cell are in metres?SDK says the span for LOD11 is 4892m - But I believe the cell is not a perfect square or is it?CheersStorm

Share this post


Link to post
Share on other sites
Help AVSIM continue to serve you!
Please donate today!

Assuming that the 4892m value specified in the SDK is the Width (horizontal length), using the Lat and Long deg boundaries for an LOD11 cell, I came up with a height of 3669.00m. Thus an LOD11 cell would be:4892m (L) x 3669m (H)Is my maths correct?Storm

Share this post


Link to post
Share on other sites

The value I calculated doesn't convince me...any ideas?

Share this post


Link to post
Share on other sites

The size depends on where on the earth you are, the closer you come to the poles the smaller they are. So your question can not be answered without knowing where on the world you are.

Share this post


Link to post
Share on other sites

Can't I use the N,S,E and W coordinates of the particular cell to calculate the span in m's in some way? These coordinates are easily computed through LOD calculator...Storm

Share this post


Link to post
Share on other sites

Yes, that should be possible. Let me think LOD11 means that the earth is split in 6144x4096 squares (I hope I counted that correct). So that would give:height = R*pi/4096width = R*cos(lat)*2*pi/6144where R is the radius of the earth, so circa 6378000 meters (if I remember correct).

Share this post


Link to post
Share on other sites

Hmm, I'm not on the same wavelength :)What I did is find the difference between the N and S coordinates of the particular LOD11 cell. Same thing for the E and W points. I then scaled these using the Lat and Lon deg values given in the SDK table for LOD11. But for this cell I came up with practically same values of 4892m x 4892m.Storm

Share this post


Link to post
Share on other sites

I noted that LOD11 cells are trapezoidal in shape with the horizontal sides parallel and of differing lengths while the sides are inclined and of the same length.I'm using degree to distance calculators to find out what the lengths are.Storm

Share this post


Link to post
Share on other sites

HI StormThey're perfectly rectangular if you if you plot them unprojected. If you plot them in any of the various projections, you'll get all kinds of irregular shapes, depending on the projection. But if you use plain old latitude/longitude, you get perfect rectangles.MSFS requires unprojected source files in WGS84 datum. If you're working with projected source data, you'll have to unproject (convert them to lat/long) them before processing for MSFS.Hope this helpsJustinhttp://www.fsgenesis.net-------Justinhttp://www.fsgenesis.netHigh Quality Scenery for FS200x

Share this post


Link to post
Share on other sites

Remember folks the earth still is a model like any other and like all3D meshes uses projections for mapping The math in the program willtake WGS48 formatted data and works out the projection based on the LOD. Thanks Justin for pointing out that the raw info must unprojected (planar) so that the textures will display correctly in the sim sice the earth is in fact a big ball(almost!) Dan

Share this post


Link to post
Share on other sites

Thanks ppl for the help, have to ask you thought to explain the jargon...things like wgs48.Anyway, what I'm interested in is the actual dimensions of say an LOD11 cell. If you say the the cells are perfectly rectangular, then all cells should be of the same size - then what size is this? and why is it someone mentioend that the cells shrink in the vicinity of the poles?If only someone could provide me with answers to:1. Are cells (say LOD11) of equal size and shape and,2. Simple way to calculate the width and lengths in m's of the cells(I tried using te degree boundaries and came up with the trapezoidal shape I mentioned before).Storm

Share this post


Link to post
Share on other sites

In degrees all cells have the same size as noted above, but in meters this means that there is a difference (because the earth is not flat). So then it is correct that you get a trapizium like thing as you mentioned.

Share this post


Link to post
Share on other sites

Cheers Arno,Ok so I'll work assuming that each cell is rectangular and all cells are of same size. I'll retrieve the size and orientation of the cells using the runways of an airport as guidlines. Should be suitable I guess.Storm

Share this post


Link to post
Share on other sites

HI Storm,You have downloaded and read the Terrain SDKs, right? Look at the Advanced Topics, Using the Level of Detail (LOD) Directive. There you will find a table with the values for all the LODs.From that table, an LOD=11 cell is19.1mWith a quadrant size of0.043945313 degrees latitude0.05859375 degrees longitude-------Justinhttp://www.fsgenesis.netHigh Quality Scenery for FS200x

Share this post


Link to post
Share on other sites

I sure read the SDK, many times over!What's the span in mtres value then? Value of 4892m ???Storm

Share this post


Link to post
Share on other sites

Hi Storm.First, we'd like to know why you would need the meter span of an LOD11 area. There would normally be little use for that.======================================The MS world is measured in degrees of latitude and longitude, not in meters. All placement of objects and effects is based on this.Some commands ( whether BGLC or SCASM ) use a distance , which is usually stated in meters, but require a world geographic refpoint. An example would be our old runway commands. This adds a lot of confusion. Gmax objects are measured in meters ( usually ), but are placed at Lat-Long coordinates.Programs like Airport for Windows, or FSSC use a span in meters for for their background bitmaps and their internal object placement ( which leads to all kinds of problems with registration ).=======================================North to South, an LOD11 span is the same in meters.... but East to West, an LOD11 span shrinks as it approaches the poles, and fattens as it approaches the equator... just as meridians of longitude meet at the poles and are farthest apart at the equator.In the real world, a meter is a meter in any direction. But when we stretch the ball-like world onto a flat 2D plane, a meter going from East to West will stretch out with great distortion near the poles in a geographic projection. Just like the distortion you'll see in many world maps ( look at Greenland ).=======================================So as Arno pointed out, the span of any LOD depends on where you are, as the LOD is a geographic measurement, not a metric distance... and conically-trapazoidal in appearance, until stretched onto a 2d plane.======================================So why do you need the span in meters?Dick

Share this post


Link to post
Share on other sites

Fair enough Dick, 10x,A fried of mine working with me on this project has got full scale contour maps/survey sheets in ACAD of the Islands. What we're trying to do is to take a top-down view of the airport runways in FS right (these should be quite accurate in terms of position and size) and match the size and orientation with the ACAD sheets. We can then proceed (hopefully) to accurately define quadrants (say LOD11) for the coastline and mesh so that orientation of coastlines (size in FS), and mesh is accurate enough.Storm

Share this post


Link to post
Share on other sites

Hi Storm.Previous posts mentioned the map projection. Your ACAD drawings are probably in equidistant projection ( everything measured in meters ). But mesh and TDF ( LWM and VTP lines and polys ) require all relative measurements in degrees of lat-long... geographic projection of WGS84 Datum.Coastline-Maker or Ground2K can use maps if they are of the right projection ( geographic ). You could use SRTM data if the islands are in the Americas, and this will be of the right projection and could serve as a base for mesh and a coastline raster image ( to use as a background image in CLM or G2K ). Your ACAD info could be redrawn onto this raster image.I use Global Mapper for most of my needs with GIS data. I could easily take ACAD data and reproject it to geographic ( simple lat-long ), and export a raster image of the drawing. Manifold is another great progrm that is even more powerful, but a bit more expensive ( I believe Justin Tyme uses this ). That drawing could even be overlayed onto a shaded representation of the mesh, to give a really great image.I believe Global Mapper, and I'm sure Manifold could take XYZ info and create a DEM off the ACAD drawing to help in the mesh creation. MicroDem might be an even better choice for this... but again, the projection must be right ( geographic / WGS84 ).Buildings would require equidistant measurement, to be built by Gmax, EOD or FSSD. These can be placed with FSSC or Airport for Windows... and they will require a background map as equidistant ( measured in meters ).Runways ( and ILS, VOR, AFD... ) are a bit of a problem for FS2004, as these are now using a new BGL format which is presently unknown. It seems to me that Microdem has the ability to create a DEM from a contour drawing. Manifold probably will do this and even Global Mapper might be able to do it.=================================Now that you have the tools, and an understanding of the need for geographic projection, you need to decide the bounds for your project.Coastline Maker and Ground2K care little about LOD boundries... these programs will just allow you to draw the roads, shorelines, streams, and watermasking to any boundry... just make sure you have geographic projection for the background image.The mesh needs geographic bounds to be resampled... and you need to have the total source bounds larger than the mesh dimensions you need... resample will trim away the "waste" from the edges. So just make sure the size of the source image is larger than the actual area you wish to mesh. Again, the DEM info needs to be in geographic projection, with WGS84 Datum.If you think you need a vertical measurement every few feet or meters for the mesh data, you are wrong. Although the vertical measurement is in meters, the x,y sampling is done from a simple division of the bounds ( which are geogrtaphic ). Say your data cover 1 degree by 1 degree. You can create a dem that splits that info into 1000 x 1000 samples... or 4096 x 4096... or 253 x 768... or any sampleing scheme you can think of. But the data points must be a division of lat-lon rather than spaced apart in so many meters. I've made up a world landclass data file that is 100 x 100 of value #254 ( transparent ) to allow me to use it in multiple source resampling, and it works fine. That just illustrates that Microsoft's Resample cares little about the "fineness" of the data, as long as it equally divides the geographic bounds of the source.Your ACAD data could be reprojected to geographic and the contour lines sampled at equal divisions. This could even be done "by hand" by gridding the reprojected image like graph paper ( equal divisions of a geographic projection ), and noting the elevation at the intersections of the graph grid.======================================Resampling of a 100 x 100 grid covering 1 x 1 degree at LOD 11 would be quite possible... the mesh would be highly oversampled. The area covered would be less than the full 1 x 1 degree size, as only the FULL LOD11 segments in the 1 x 1 degree area are resampled... the edge material is trimmed as waste.More points ( equal divisions of the source data ) = more accurate mesh. If you are trying to get the most accurate LOD 11 resampling, the data will require a minimum of 256 divisions per LOD11 Area. More than this number will just make Resample work a little harder, but is not a bad thing.256 is the magic number. Each LOD size, whether LOD5, 6,7,8,9,10,11... will use 256. Resample will add a point to the end result so a division of one LOD would have 257 points of data... 2 adjoining LODs would have 513 points... 3 would have 769 points. You don't need to worry about that as Resample will take care of it. All you need is as many equally-spaced geographic divisions as you can add... but you don't need to go over the 256 per LOD, as that will be resampled away.An LOD11 area is 0.9375* x 0.703125* ( Long x Lat )... so the 256 points maximum resampled spacing is 0.003662109375* x 0.00274658203125* ( Long x Lat... or x,y ). That's the grid spacing in terms of degrees. You can use closer spacing, but Resample will adjust the resultant mesh... just as it will adjust farther spacing ( guessing what to fill in ). You can use different spacing... for example, equal spacing ( 0.003 degrees spacing would be close for LOD11... 0.0025 x 0.0025 would be better, without too much waste ).Resample does a good job of forcing the data to the 256x256 matrix.=========================Another approach that might work is the grises50 program. That will allow you to create a simple 8-bit raster image of the area and convert it to a mesh. The package includes a good tutorial.Spain Mesh Group:http://www.simuvuelo.org/pags06/tutorial.htmIf you brush up on your Spanish, this site has a goldmine of info on creating mesh. Dick

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this