Hi,

I am doing some work with intersecting layers at present, and getting some quirky results when analysing the areas from the splits.

 

When I intersect a paddock with a soil layer, I calculate the area of the resultant intersected areas.

I usdate the result with the original paddock ID and the original Paddock area (as Floats).

In theory, the sum of the resultant areas from each paddock should equal the total area of the original paddock.

I also calculate the proportion of the original paddock for each resultant area (result area/original area).  Sum these up for each paddock and you should get 1.

If I use the function Area(obj,”hectare”), the sum of the proportion of the areas ranges from 0.990 to 1.01 with a few more extreme outliers.

I get better results from using CartesianAreas(Obj,”hectare”), but can only use this for non lat/long projections, but actually get mainly 0.996 as the sum of the proportions (which is similar to the AMG factor of 0.9996)

I have read some of the posts about this a while ago (although they were mainly concerned with distance), and they suggested a reasons for this being the scaling factor used in the projections – although I didn’t quite understand it.

I have some questions:

Is there something I should do to get a better result?

Can I test for the projection being a lat/ long or not easily? Or should I use a function that does something like take the projection string and search for the word latitude?

Is anyone else worried about how inaccurate these values are?

 

Thanks

R

 

-------------------------------------------

Robert Crossley

Agtrix P/L Australia

 

Far Southern Queensland Office:

9 Short Street

PO Box 63

New Brighton 2483

 

P: 61 2 6680 1309

F: 61 2 6680 5214

E: [EMAIL PROTECTED]

W: www.agtrix.com

 

Brisbane Office:

109 Milsom St

Cooparoo  4151

Queensland

P: 61 7 3843 3363

 

 

 

 

_______________________________________________
MapInfo-L mailing list
[email protected]
http://www.directionsmag.com/mailman/listinfo/mapinfo-l

Reply via email to