# Locus Map - forum

## Support => Troubles & Questions => Topic started by: balloni55 on June 09, 2017, 14:16:08

Title: UTM projection is not precise
Post by: balloni55 on June 09, 2017, 14:16:08
Hi menion
why are the coords of a projection not accurate?
i have done example projection 5000m N/E/S/W
horizontal projection isn´t accurate in the x-axis
vertical projection isn´t accurate in the y-axis
(https://s11.postimg.org/twrjngqsv/image.jpg) (https://postimg.org/image/twrjngqsv/)
(https://s2.postimg.org/g9ah5jvlx/image.jpg) (https://postimg.org/image/g9ah5jvlx/)
Title: Re: UTM projection is not precise
Post by: balloni55 on June 16, 2017, 13:07:46
Hello menion
............ :P
Title: Re: UTM projection is not precise
Post by: menion on June 16, 2017, 14:14:22
Hi :)
I saw this topic in metro and wasn't sure what is wrong with these coordinates. And when I look on them now, I still do not know.

Aren't they correct? On first image, second values are same ( longitude ) and first value ( latitude ) differ, which is correct, no?

Also did you tried to use some web convertor like http://www.rcn.montana.edu/resources/converter.aspx , to check if computed UTM coordinates by Locus Map are correct?
Title: Re: UTM projection is not precise
Post by: balloni55 on June 17, 2017, 19:04:31
Hello menion
UTM coords are in meters, correct?
and as i know it´s possible simply add or subtract distance meters in x-axis (E/W) and/or y-axis  (N/S) to get the projected coord.
Actually locus calculate too short distances if projection directly point to 0°/90°/180°/270°
intermediate values i didn´t know

if projection distance is:
- 2500 m, result is 1m too less
- 5000 m, result is 2m too less
- 10000 m, result is 4m too less
so in my screenshot example with 5000 meter projection, locus calculate only 4998 m :-[

I hope it´s now more clear ;)
Title: Re: UTM projection is not precise
Post by: balloni55 on June 28, 2017, 09:14:46
Hello menion
is my description still unclear?
or is my knowledge wrong how to calculate UTM??
Title: Re: UTM projection is not precise
Post by: menion on June 28, 2017, 09:30:18
Nono all is clear, problem (delay) is on my side. I have so many tasks to do, that on this, quite un-important for me now, I always forget ... sorry
Title: Re: UTM projection is not precise
Post by: sonny on June 28, 2017, 18:32:24
I think Locus is treating UTM-coordinates correct. But for checking with professional conversion software you should provide the coordinates in degree of your points too.

In general a UTM- grid is not exactly "rectangular" compared to the degree-grid of the same region. But it is more or less rotated, meaning the meridian of the UTM-grid are *not* running from geographical North to south. Please take a look at the image:
(https://s23.postimg.org/peizy0ggn/d_utm.jpg) (https://postimg.org/image/peizy0ggn/)
As you can see the black lines of degree-grid are not parallel but lightly rotated to the colored UTM-grid, especially on the sides of an UTM-zone.

The other point to consider is, that UTM is just a projection of the earth's surface which is nearly a sphere and can't 1:1 projected into a rectangular system. So simplified: 1 meter in UTM grid is just "at average" 1 m in reality, depending in which location within an UTM-zone you are measuring. The differeces are of course very low, but a distance on real earth of 5000 m could be a just distance of for example 4998 m in UTM-grid.

So we can't 1:1 convert UTM-meters in real-world meters. Also each grid system uses another model of Earth with slightly different diameters of their models.
Title: Re: UTM projection is not precise
Post by: menion on July 02, 2017, 23:08:41
Good day guys,
sonny, thank you for a very precise answer a lot better that I personally should give. There is nothing I may say more, thanks!
Maybe an interesting fact for @balloni55. Distance error is not linear. It grows based on distance from a center of every UTM Grid anyway it grows slowly at start but it increase a lot (with square) when distance since center meridian grows.
You may try to compute half distance, only 2.500 m and error should be close to quarter.