TMC file coordinates format

    Quote from drhotwing1

    Trespassers

    What exactly are you having trouble figuring out? Maybe you are just explaining this in a way that I'm not understanding. We are just trying to help you out here.

    Thank you Jeff.

    I try to understand the exact technical background behind the TMC coordinates and how they're actually being used by GeoConvert tool, not just repeating woodoo rituals.

    Let's try summarizing what we have so far:

    A) From your post #3 I read there might be a kind of bug causing Geoconvert tool to sometimes fail for (yet) unknown reasons when more decimals are given in the TMC file... I never experienced it yet, but it is well possible.

    I can live for now with rounding up the coordinates in the TMC file, what I definitely need then is to know how these rounded up coordinates are being used by Geoconvert in order to set them accordingly. Therefore my request for clear rounding up guidance in order to feed the tool with appropriate values in all cases.

    B) From Phil's post #5 and your post #6 I read that the coordinates in the TMC should be larger than the scenery zone to compile.This could be understood as an envelope, and Geoconvert tool will compile any found picture having an AIC file with coordinates inside this envelope.

    Looking at Hartman's post #17, it would mean the TMC file should always be the blue case.

    If this was really the case, Hartman's yellow case in post #17 would compile nothing.

    C) From Rodeo's post #18 I read however that both Hartman's blue and yellow cases are possible with the same orthophoto (and same AIC file, thus).

    If Rodeo is right, it means B) is not correct.

    What is then being compiled in Hartman's yellow case?

    - The complete picture ?

    - Only the yellow square of the bigger orthophoto will be compiled?

    If only the yellow part of the photo is being compiled, it means the TMC coordinates "crop" the orthophotos (building a blending mask outside ?), which may be useful in some cases, but the 0.00° coordinates format is definitely coarse.

    If it's the case, what happens exactly in Hartman's blue case? is the blue part replaced with a blending mask?

    I know it will take time to build complete Wiki documentation, therefore in the meantime if we can have some clarification on the forum that would be very helpful...

    Thanks in advance

    Cheers

    Antoine

    Config : i7 6900K - 20MB currently set at 3.20GHz, Cooling Noctua NH-U14S, Motherboard ASUS Rampage V Extreme U3.1, RAM HyperX Savage Black Edition 16GB DDR4 3000 MHz, Graphic Card Gigabyte GeForce GTX 1080 8GB, Power supply Corsair RM Series 850W, Windows 10 64 bit.

    Edited once, last by Trespassers (July 20, 2017 at 10:14 AM).

    Hi,

    together we're approaching a better comprehension of geoconvert.

    Let me summarize my understanding:

    For each aerial image we need a coordinate file. This specifies the geoposition and extension of the image in a global view.

    One image can only have one coordinate file.

    (Generally: It's a fact that there exist many different coordinate systems, projections, ellipsoidparameters.
    They derive from the task to project a globe area onto a flat map. Dependant from the task and the position on earth equal of angle projections or true aera projections are used. aerofly can handle different setting, I know at least lon_lat and webmercator coordinate systems.)

    The TMC lonlat_min and lonlat_max comprise an area for geoconvert.
    It does not matter whether this area consists of a single aerial image or several images.

    Since the TMC lonlat_min and lonlat_max can be set separately for different levels,
    you can define a larger area for low levels (9,11) and a smaller area for higher levels (12-14).

    Again, it does not matter whether this area consists of a single aerial image or several images.

    aerofly FS 2 has it's built in tile system.

    It converts areas dependent from TMC lonlat_min and lonlat_max and level to the specific aerofly tiles.

    Geoconvert will run to compile the full aerial image (blue) for the lower levels, then compile only the inner yellow area for the higher levels.

    Blending will take place, where an aerofly tile is cut by the TMC lonlat_min and lonlat_max.

    So it is not necessary to blend the complete outer blue area of Hartman's image, only sections to finish a full aerofly tile.

    I appreciate this discussion to make things clearer for all of us interested users.

    Any comments are welcome.

    Regards

    Rodeo

    Okay, now it's clear, thank you so much!

    The last point is to understand why and in which case decimals further than 0.00° would cause GeoConvert to fail.

    Sub-pixel decimals will sure cause troubles, but as long as we match pixel size it should work. To be investigated.

    And, of course, I also very much appreciate this constructive discussion, thanks for sharing!

    Cheers

    Antoine

    Config : i7 6900K - 20MB currently set at 3.20GHz, Cooling Noctua NH-U14S, Motherboard ASUS Rampage V Extreme U3.1, RAM HyperX Savage Black Edition 16GB DDR4 3000 MHz, Graphic Card Gigabyte GeForce GTX 1080 8GB, Power supply Corsair RM Series 850W, Windows 10 64 bit.

    Thanks Antoine,

    I forgot to mention another option.

    If you do an inner area for higher levels from the same source (one or more aerial images),

    you may get masks where the internal aerofly tiles are cut by the TMC frame (this is what I explained before).

    But since you have a consistent image source you can remove these ..._mask.ttx files to increase performance.

    aerofly then displays the full highres tile instead masking part of it.

    Hope this is a comprehensible explanation.

    Rodeo

    It is, good input, thank you !

    cheers

    Antoine

    Config : i7 6900K - 20MB currently set at 3.20GHz, Cooling Noctua NH-U14S, Motherboard ASUS Rampage V Extreme U3.1, RAM HyperX Savage Black Edition 16GB DDR4 3000 MHz, Graphic Card Gigabyte GeForce GTX 1080 8GB, Power supply Corsair RM Series 850W, Windows 10 64 bit.

    • Official Post

    Thank you everyone for contributing to both the questions and answers. It really helps out users that don't quite yet understand the process.

    I will review all of these questions and answers then will be moving some of it over to the wiki for future review.

    It's great to see questions like this because it shows that this tool is that important to the community :)

    I think that the phrase of the week should certainty be "Have you Geoconverted today?" Maybe it will catch on.

    IPACS Development Team Member

    I'm just a cook, I don't own the restaurant.
    On behalf of Torsten, Marc, and the rest of the IPACS team, we would all like to thank you for your continued support.

    Regards,

    Jeff

    I would appreciate a discussion in the wiki concerning the "source" files. I have found the USGS aerial images to be the most recent or most up- to-date in south central U.S. The Virtual Earth images seem to be availabe for a much larger global area but, are woefully out- of-date.

    The Google images seem to be the most realistic and most pleasing, probably due to both being somewhat recent and and being taken in the spring and summer. The VE images are both old and taken in the winter when all the grass is dead and many trees do not have leaves.

    I'm sure all this will vary with specific locations and other factors. Bing is not available in my area.

    A comparison image would be nice. One area, all sources compared.

    Regards,

    Ray

    Quote from Jetjockey10

    A comparison image would be nice. One area, all sources compared.

    As you say, it very much varies from place to place.

    There's no clear best provider, that's why for FSET many "services" can be set, and when trying to build up a scenery, you simply browse among your various services, looking for the one locally providing the best pictures.

    Google by the past added a captcha to prevent robots (FSET is to be understood as a robot) from grabbing images.

    Roland had to add a popup window where you could type the captcha to resume process, but then you had to stay at your computer all during the grabbing and re-enter a captcha each few dozens of tiles...

    Virtual Earth has globally rather good pictures in Europe, but some places are really poor.

    BTW it is also the case when legally purchasing aerial photos from official agencies and some regions need a huge amount of manual work...

    For the US, it seems USGS comes on top, not only in raw material quality, but also it is the only service yet to allow this usage AFAIK.

    Cheers

    Antoine

    Config : i7 6900K - 20MB currently set at 3.20GHz, Cooling Noctua NH-U14S, Motherboard ASUS Rampage V Extreme U3.1, RAM HyperX Savage Black Edition 16GB DDR4 3000 MHz, Graphic Card Gigabyte GeForce GTX 1080 8GB, Power supply Corsair RM Series 850W, Windows 10 64 bit.

    Edited once, last by Trespassers (July 20, 2017 at 3:31 PM).

    Quote from Jetjockey10

    Hmmm. I missed seeing the Google logo. I will make a closer inspection. Thanks.

    Regards,

    Ray

    You'll see it better if you zoom in water areas (lakes, for instance).

    Config : i7 6900K - 20MB currently set at 3.20GHz, Cooling Noctua NH-U14S, Motherboard ASUS Rampage V Extreme U3.1, RAM HyperX Savage Black Edition 16GB DDR4 3000 MHz, Graphic Card Gigabyte GeForce GTX 1080 8GB, Power supply Corsair RM Series 850W, Windows 10 64 bit.

    New issues - new solutions?

    In post #17 I asked:

    "Do you mean that we can, for example, also choose both blue (outside) and yellow (inside) area from the same orthophoto?"

    I have some orthophoto that I want to convert (Note: Not from FSET).

    That means I have to create many .tfw files manually.

    See illustration.

    Is it possible to create a single .tfw file for blue area, and then use this file for all the "green" photos inside?

    Blue.tfw (blue area = source).

    Green1.tfw ("green" photo 1 - same text content as blue)

    Green2.tfw ("green" photo 2 - same text content as blue)

    Green3.tfw ("green" photo 3 - same text contentas blue)

    Green4.tfw ("green" photo 4 - same text content as blue)

    .... etc.

    Will all "green" photos within the blue area be correctly converted?

    Edit: This was a wish dream. I think my brain is boiling over now ... :sleeping:

    nope, 1 TFW for each photo.

    But 1 TMC for the whole set.

    Cheers

    Antoine

    Config : i7 6900K - 20MB currently set at 3.20GHz, Cooling Noctua NH-U14S, Motherboard ASUS Rampage V Extreme U3.1, RAM HyperX Savage Black Edition 16GB DDR4 3000 MHz, Graphic Card Gigabyte GeForce GTX 1080 8GB, Power supply Corsair RM Series 850W, Windows 10 64 bit.

    Quote from Hartman

    Thanks for confirmation.

    That's the reason I like TMC better than TFW. :)

    If you need a tool to create tfw for you i can have that to you by tomorrow, I have one that takes ~25 seconds to process a JP2 downloaded from the usgs into a TIF with TFW.. I am also working on the bit that defines the total TMC area for the dataset to be processed. program loops through a working directory and deposits them in a "done folder".... Maybe tonight since the wife is headed out to meet friends.

    You need to install osgeo4w.. I will create the exe.. you need to be able to modify a config file to point at the directories you are working with. (Of course the program is open source)

    Quote from Sycosys

    If you need a tool to create tfw for you i can have that to you by tomorrow, I have one that takes ~25 seconds to process a JP2 downloaded from the usgs into a TIF with TFW.. I am also working on the bit that defines the total TMC area for the dataset to be processed. program loops through a working directory and deposits them in a "done folder".

    That's amazing. :thumbup:

    I already have some great orthophotos on a hard drive (as I've got on DVD discs). This is a lot of work to convert, as I have to do most of the preparations manually. I therefore try to find shortcuts.

    Quote from Sycosys

    Maybe tonight since the wife is headed out to meet friends.

    Hope she will enjoy herself and stay out for a long time..... ;)

    Quote from Jetjockey10

    Hmmm. I missed seeing the Google logo. I will make a closer inspection. Thanks.

    Regards,

    Ray

    Poo. All the beach water has a white sheet covering it with a logo on it. So much for using Google scenery. Have to see how the USGS handles the water.

    Regards,

    Ray

    Don't give up on the water issue. There are other threads showing that if you use the editor called gimp or equivalent that can make the water A transparent area called Alpha, then it will blend with the water that already exists in the lo-res world