2

I have a long route (~2400 KM) that span 3 different UTM zones. (N - 37, 38, 39)

On this route there is a sign every 1 km-approximately (km marker), these signs help in determining the exact location of assets along the road etc…

I have the actual location of these signs as a point and it turned out that the distance between 2 markers is between 1100 m to 1200 m in general so there is an error in these markers location caused by human mistakes. So I need to create new km markers that conserve the distance between the 2 markers (1000 m) using linear referencing and event tables on my road route.

But the problem is my route passes by 3 different UTM zones and all of my data are geographic WGS84 which is in degree, therefore, I can’t create km markers using linear referencing and event table in this case for each 1 km as my m measurements if the route was WGS 84.

And I need the rout to be 1 single record along the 3 UTM zones so I can’t divide the route based on the projection boundary. Even though I tried to do this and also had an error equal to 20 m on the boundaries of the projection zone.

What should my route projection be in this case? So that the linear referencing and event tables work without any distortion.

For sure the route should be in a projected coordinate system since we're using meters (1000 m) to dived the route or you think we can use the geographic coordinate system by transforming the m value to a decimal degree.

2

5 Answers 5

1

The definite solution would require adopting a quasi-linear map projection such as the snake projection.

0

I'd suggest a two-point equidistant projection.

2
  • thank you felipe, actually, I did try the two-point equidistant and it didn't work well, on the curve the distance between my physical point and the point created by the event table will increase. and that happens only on curves. Commented Oct 25, 2018 at 6:51
  • what is the mean altitude in the region of interest? apart from the projection, you might also need a ground-to-ellipsoid distance reduction. Commented Oct 28, 2018 at 13:29
0

I think the m value of the vertice of a polyline are independent of the projection.

So if I'm not wrong you should be able to let your data in WGS84 (degree) and reference it with m value in meters (or km), the difficulty being to attribute accurate m value to each vertices...

EDIT : The m value has no real unit or real geographical meaning, if you set the first vertice m=0 and the least vertice m=10, then a m value of 5 will be at the middle of the line regardless of the coordinate system unit or the actual length of the line.

So you could let your polyline in any GCS you want, you just need to have some point located on the line for witch you know the exact km from start and you could use these point to calibrate your line.

1
  • thank jr, you mean the polyline should be also WGS 84? but in this case, how can I set the m value in km? mean how can I reference it with value in m Commented Oct 24, 2018 at 10:05
0

You can try designing a low distortion projection (LDP):

https://www.plso.org/resources/documents/dennis%20ground_truth_handout_v22_plso_2015.pdf

0

Based on your comments I came up with the following solution:

Basically I created a route, chose the predominant UTM zone and built it using that projection. The route itself is "Relative distance along its length" -- so as long as your "perceived" measured length matches the Route Length Measure all events along the route will be proportionate.

I say "perceived" because you can stretch and lie to the route about its dimensions to fit along with the physical or political measurements. I always set up calibration points at the transitions between UTM zones to account for any differences due to the projections.

I had some physical point on the track that I know its location of so based on these points I did the calibration of my route.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.