[Bf-python] a geoToUTM() function in a bpy module ?

[jerome]

oh I meant no time for the proposal process, if you agree for sure I can push a geospatial_utils.py module with geo_to_utm() in it.

such module would need a real mathematician to improve and to address common geospatial needs, if its only me it will turn into a 2 functions module I won't be able to explain exactly ;)

Jerome

Le 10/09/2012 02:14, Campbell Barton a écrit :
> Hi jerome, such functionality is indeed handy, but think it belongs in
> its own module,
> perhaps "bpy_extras.geospatial_utils"
>
> ./release/scripts/modules/bpy_extras/geospatial_utils.py
>
> Also, wouldn't it make sense to have a function to do the reverse conversion?,
> geo_to_utm() / utm_to_geo()
>
> If you don't have time to do this, its fine, but could add in a
> comment that its a  TODO.
>
> On Mon, Sep 10, 2012 at 7:16 AM, jerome <jerome.le.chat at free.fr> wrote:
>> Hello,
>>
>> I'm currently programming things about city generation for a BGE project
>> I have.
>> open street map is a really valuable input for such need, as you know I
>> suppose,
>> since you can retrieve a lot about city geometry worldwide, and generate
>> from it in Blender.
>>
>> in a recent commit I updated a bit the osm importer to add a better
>> projection from lat/lon to blender units.
>>
>> I think this function, or an equivalent one, should be part of the bpy,
>> maybe in mathutils.geometry, or a more suitable location as you wish :
>>
>> it's really a multi usage function.
>> this could help to bridge with the osm community,
>> and with architects too.. for now the ones I know are a bit reluctant to
>> Blender but I'm hardly working on it, BGE helps a lot actually.
>>
>> the tests I'm doing with lxml xml parser are very conclusive to read
>> write huge osm or extended osm quickly.
>> by extended I mean extra tags about height, uvs,utm coords. a kind of
>> .bosm format I'm writing.
>>
>> anyway here's the proposed function, consider it copyleft.
>> sorry if my proposal does not respect blender guidelines, but I really
>> have no time left :s
>>
>> regards,
>>
>> Jerome / littleneo
>>
>>
>> (from math import radians, sin, cos, tan, sqrt)
>>
>> # given lat and longitude in degrees, returns x and y in UTM (1 KM = 1
>> BU ) .
>> # accuracy : supposed to be centimeter. community feedback needed.
>> # looks ok so far
>> # http://fr.wikipedia.org/wiki/Projection_UTM
>> # http://fr.wikipedia.org/wiki/WGS_84
>> # http://earth-info.nga.mil/GandG/publications/tr8350.2/wgs84fin.pdf
>> #
>> http://geodesie.ign.fr/contenu/fichiers/documentation/algorithmes/alg0071.pdf
>> # wiki is your friend (don't ask me about math Im just a writing monkey.)
>> # jerome.le.chat at free.fr
>> def geoToUTM(lon, lat) :
>>
>>       # if abs(lat) > 80 : lat = 80 #wrong coords.
>>
>>       # UTM zone, longitude origin, then lat lon in radians
>>       z = int( (lon + 180)  / 6 ) + 1
>>       lon0 = radians(6*z - 183)
>>       lat = radians(lat)
>>       lon = radians(lon)
>>
>>       # CONSTANTS (see refs.)
>>       # rayon de la terre à l'équateur
>>       a = 6378.137
>>       K0 = 0.9996
>>       # flattening consts
>>       f  = 0.0033528106647474805  # 1 / 298.257223563
>>       e2 = 0.0066943799901413165  # 2*f - f**2
>>       e4 = 4.481472345240445e-05  # e2**2
>>       e6 = 3.0000678794349315e-07 # e2**3
>>
>>       # lat0. 10000 for South, 0 for North
>>       N0 = 10000 if lat < 0 else 0
>>
>>       A = (lon - lon0) * cos(lat)
>>       C = (e2 / (1 - e2)) * cos(lat)**2
>>       T = tan(lat)**2
>>       vlat = 1 / sqrt( 1 - e2 * sin(lat)**2 )
>>       slat = (1-(e2/4)-((3*e4)/64)-((5*e6)/256))*lat -
>> (((3*e2)/8)+((3*e4)/32)+((45*e6)/1024))*sin(lat*2) + (((15*e4)/256) +
>> ((45*e6)/1024) )*sin(lat*4) - ((35*e6)/3072)*sin(lat*6)
>>       E = 500 + (K0 * a * vlat) * (A + (1-T+C)*((A**3)/6) + (5 - 18 * T +
>> T**2) * ((A**5)/120) )
>>       N = N0 + (K0 * a) * ( slat+vlat*tan(lat)* (A**2/2 +
>> (5-T+9*C+4*C**2) * (A**4/24) + (61-58*T+T**2) * A**6/720) )
>>       return E,N
>> _______________________________________________
>> Bf-python mailing list
>> Bf-python at blender.org
>> http://lists.blender.org/mailman/listinfo/bf-python
>
>