Recreating real terrain from Goole Earth

Emx77

Recruit
Joined
May 7, 2008
Messages
12
Reaction score
0
Location
Sarajevo
Country
llBosnia
Recreating real terrain from Google Earth

I'm trying to recreate real terrain map but I have faced some difficulties. My goal is to use Google Earth and Google Terrain overlay in combination with AATF Power Toolkit (AATF PT) to recreate real terrain. I decided to go with 6 x 5 km map. I followed "Creating Maps from SRTM Data - Tutorial" and sucessfuly obtained XDM map with size of 1 degree by 1 degree square where bottom left corner coordinates are 43 deg. 47' 10" LAT and 18 deg. 58' 26" LONG. Next step is creation of elevation map in AATF PT and that is where real problems start.

1) When "Map Production Information" dialog is shown if I accept all default values and simply clik "OK" button, I get 50 x 50 km map. So far so good. But, as I already stated, I want smaller 6 x 5 km map.

After adjusting map dimension in "Map Product Information" dialog, the resulting smaller map is distorted along upper and right edge. Also some of contours are disconnected.



I am not sure why this disstortion appears? Is it a bug or maybe I need to change some of the default parameters in Map production Information" dialog?

2) Resulting map of 6 x 5 km have 2400 x 2000 pixels. Doing simple math I concluded that 1 km = 400 pixels. After adjusting Google terrain map (which have 20 m contour lines) to have same 1 km = 400 pixels ratio, I would expect to get exact match after I overlay Google terrain map over AATF contour map in drawing program, but it is not a case. They are comparable but not (exactly) matching. I assume solution could be in adjusting map construction parameters in AATF PT "Map Production Information" dialog, but I don't know how to obtain these parameters from Google Earth.

3) After I convert my map bottom left corner coordineates of 43 47' 10" LAT and 18 58' 26" LONG in seconds I get 157630 LAT and 68306 LONG which does not correspond with 154800 LONG and 64800 LAT which are default values in "Map production information" dialog after opening XDM file. I wonder why these values are not matching and which values should I enter?

4) Having all of the above in mind, does anybody know how to recreate real terrain using Google Earth and AATF PT?

Many thanks in advance.
 
Last edited:

Pat Proctor

President, ProSim Company
Joined
Oct 5, 2002
Messages
1,189
Reaction score
1
Location
USA
Country
llUnited States
The default values in the map dialog are the same, no matter what DEM you load, ith the exception of the lat long values. The lat-long values should be the bottom left of YOUR DEM. You will have to change them to the bottom left of the map you are trying to create (which should always be a higher number than the default values). If the default values for lat long in the map dialog do not match what you think should be the bottom left lat long of the DEM you created using SRTM data, you might need to recreate the DEM.

To get a map that EXACTLY replicates the terrain on Google Earth, you will need the m/s and alpha angle for the map section.

The m/s entry is, literally, the number of meters in every arcsecond of longitude. The Alpha number is the difference between true north and grid north in mils (6400 mils per 360 degrees).

You can determine this data in a number of ways, but the best way is by loading the USGS DEM in MicroDem. This is a free download available here:

http://www.usna.edu/Users/oceano/pguth/website/microdem.htm

Use the info function to get the grid transform angle. That angle will be in degrees, which you can convert to mils. To figure out the m/s, you may have to divide the total width of the map in meters by the total width of the map in arcseconds (3600 if you have a 1 degree by 1 degree DEM).

There is nothing to be done about the edge effects on the bitmap. It is a quirk of the drawing program.
 

Emx77

Recruit
Joined
May 7, 2008
Messages
12
Reaction score
0
Location
Sarajevo
Country
llBosnia
Pat, I would like to thank you for realy useful information. During, last two days I have been trying to use all of above information in order to recreate small portion of terrain. Unfortunately, I haven't succeeded yet as I am probably doing something wrong.

Use the info function to get the grid transform angle. That angle will be in degrees, which you can convert to mils.
In MicroDem function "Map information" gives following:

True north at top of map
DEM native datum: WGS72
--------------------------------
Magnetic declination: 3.5°
UTM Grid declination: -1.2°

At the same time function "Horizontal Earth Curvature" provide:

Point: N43°46' 0.01" E18°58' 0.00"
Grid true angle: -1.56

I noticed that -1.2° (-21.3 mils) is some kind of average for whole map area, while -1.56 (-27,7 mils) is value for bottom left corner of map. Also, with function "UTM-grid true north declination" it is possible to get value for exact spot on map (for example -1.54). I hope that these are "grid transform angle" values you have mentioned before. If so, I presume that best measure is one obtained through "UTM-grid true north declination" for map area I want to recreate. Right?

Nevertheless, I tried all of mentioned values on my map but resulting AATF contour map doesn't match Google terrrain map and there is always some discrepancy. If I entered right coordinates, mils and m/s, could there be any other reason for this discrepancy? Different map projections maybe? Also I didn't change default 8-digit UTM values and Grid Zone Designators in Map Information Dialog (as I presume that is enough to put right long. and lat. coordinates).
 

Pat Proctor

President, ProSim Company
Joined
Oct 5, 2002
Messages
1,189
Reaction score
1
Location
USA
Country
llUnited States
You are doing the transform angle correctly.

The AATF map will never EXACTLY match the google map. You are taking a map with data at about 100m x 75m intervals and rotating and interpolating elevations on different axes at exactly 100m x 100m intervals. There will be some unavoidable rounding errors.

That being said, how large is the discrepancy? Is it across the entire map? If, for instance, the map is consistently 100m x 370m to the west and north of the data in google, you may be able to just "fudge" the numbers in the map creation dialog to make the two maps consistent. I have seen this happen before in our own maps (the falklands) between paper and DEM maps. The problem had something to do with datums, and we fixed it with this simple workaround.

If the map is completely different than the google map, there may be a more dramatic problem, like a DEM being extracted incorrectly from the SRTM data by the online program.

If you could give me screenshots showing the discrepancy in a specific area, I may be able to walk you through some workaround.
 

Emx77

Recruit
Joined
May 7, 2008
Messages
12
Reaction score
0
Location
Sarajevo
Country
llBosnia
Ok, I will try to describe whole process.

1) I created XDM map with bottom left corner coordinates N43°46' 0.00" E18°58' 0.00".

2) In Map Production Information (MPI) Dialog I entered following:

Long. E18°58' 26.00" (68306")
Lat. N43°47' 10.00" (1757630")

Alpha (mils): -27.4
m/sec: 22.33

Dimensions: 6.0 km x 3.9 km

In this way I obtained AATF contour map with bottom left corner coordinates E18°58' 26.00" N43°47' 10.00".

3) I used Google Earth with Terrain overlay and added placemark at E18°58' 26.00" N43°47' 10.00" (same coordinates as coordinates I have entered in AATF MPI Dialog), and adjusted GE map to have 2.5m/pixel ratio as AATF map.

4) In drawing program (Paint.NET) I overlayed GE map over AATF map.

5) Then, I matched placemark coordinates in GE map with bottom left corner of AATF map which should have same coordinates as placemark. Result is shown at "Image 01". I marked some of dicrepancies with blue lines. As you can see, these discrepancies are pretty high and it seems that they getting smaller as we move from north to south.

6) In the next step I panned GE map to match, as closely as possible, long contour east of the road. Resulting "Image 02" had much smaller discrepancies but their direction implied that GE map overlay shoud be rotated in clockwise direction.

7) Finaly, I rotated GE map layer for 2°and resulting "Image 03" is pretty close match.

This is closest I can get recreating GE map over AATF map. I wonder if this is close enough or it could be done something better to improve result (or maybe simplify whole process)? For example, I'm puzzled why such big disrepancy in "Image 01" if GE placemark coordinates and bottom left coordinates of AATF map are the same?


LINKS for images:

Image 01 - http://img26.picoodle.com/img/img26/3/12/30/f_Image01m_86ecbbf.gif
Image 02 - http://img37.picoodle.com/img/img37/3/12/30/f_Image02m_d78e43b.gif
Image 03 - http://img34.picoodle.com/img/img34/3/12/30/f_Image03m_3c623e8.gif

or Rapidshare link:

http://rapidshare.com/files/178119465/Images.zip.html
 

Pat Proctor

President, ProSim Company
Joined
Oct 5, 2002
Messages
1,189
Reaction score
1
Location
USA
Country
llUnited States
Ahhhh... Now I understand what you are saying.

First, the GE map is ALWAYS going to be rotated with respect to the AATF map. That is what the alpha angle does. The GE map is oriented with north as true north. The AATF map, like a military map, is oriented with grid north as north. If you are seeing a 2 degree difference, this is roughly consistent with the ~1.5 degree alpha angle you have seen.

As for the panning, this is the same problem I described with the Falkands maps we built. It is some kind of datum issue in how the elevation data is recorded. The error is visible, but not dramatic (less than 500m in each direction). If you want your AATF map to fit the GE map exactly, you can adjust the bottom left of your map by a few arc-seconds in order to "pan" the AATF map onto the GE map.

The more fundamental question, which data is right, I cannot answer. I learned exactly enough about Geospatial science to build AATF and the scenario builder. But, ultimately, any map is just a representation of actual terrain. A datum, if I understand correctly, is just the algorithm used to convert the round earth into flat maps. None is more correct than any other. the SRTM data, I believe, is in WGS 84. I don't know what GE maps use. But I think that this difference is a result of different datums. By "fudging" the arcseconds in your AATF map, you are just aligning your data to the GE datum.
 
Top