Resection error

It depends.... what program are you using? You should have seen deltas/ residuals while collecting the data for your resection. ....2 pt resection is by no means optimal... That should be similar to what youƒ??re seeing when you occupy another point.?ÿ

0.05 at random, rotated in the same direction, different radius from resection point, or what?

Can you shoot from one control point to another (even if you can't see everything else) to check the original distance versus yours?

1. Have you verified that the original control points are in stable condition and have not moved or settled since the last time they had measured?

2. How many points have you resected off? Generally my rule of thumb is to use a minimum of 3 points to get a clearer sense of the error residuals. You are able to get a solution using only 2 points, however that does not guarantee a check against any problems in the original point.?ÿ

3. I believe most total station programme for resection should be robust enough if it is able to calculate both angles and distances.?ÿ

What was the angle between the control points??ÿ ?ÿflat angle turns between control points is going to yield larger errors.?ÿ I usually try to resection closer to 90?ø

You haven't given us a lot of details. But 0.05' error is completely plausible.?ÿ I insist that resections be done with at least 3 points, and it is best if the instrument and the layout area is inside the triangle formed by the 3 control points. It is generally good if you can stay within an area covered by that triangle flipped along one edge. If that makes sense.?ÿ Get much further and it gets more iffy.?ÿ

3 point resections yield some form of residuals and positional error estimate.?ÿ ?ÿ

Posted by: @va-ls-2867

?ÿflat angle turns between control points is going to yield larger errors.?ÿ I usually try to resection closer to 90?ø

This is a common misconception, common enough that I had it before being challenged on it and investigating some years ago.?ÿ With a 2-point resection using modern instruments, you'll actually get a slightly smaller error ellipse when your resection point is on a line between the control points than if the points are 90?ø apart.

Posted by: @jim-frame

The idea comes from angle-only resections.?ÿ When you measure both angle and distance the old limitations are relaxed.?ÿ You still need control points spread around the work area, but being on line between two points is not a problem.?ÿ

If you have a least squares program you can investigate possible layouts to get a fuller understanding.

Resection and strength of figure concepts have significantly evolved with technology- the professional knows why heƒ??s doing what he does, what his error budget is, how to manage it, how much redundancy is enough.?ÿ

Sometimes, if a traverse adjustment is made, you are resecting to theoretical locations.

The coordinates may have changed, but the points in the field did not.

Plus, the flatter the angle, the greater the uncertainty.

AJF

Purely angle resection has a rich history and can be fraught?ÿ with uncertainty depending on the angle(s) from control to the target station.?ÿ *But*, adding accurate EDM observations these days has changed the scenario to where the target coordinates are close to the accuracy inherent in the control network no matter the geometric configuration.?ÿ So when questioned about resection observations one must first ask "were distance observations part of the system?"?ÿ If so, it's no longer a resection, instead it's a combo triangulation/trilateration observation which is mathematically a different breed of cat.?ÿ Don't know what to call it.

@mike-marks Some people, particularly Europeans, call it Free Stationing.

An interesting analysis is in the paper "Optimum Establishment of Total Station," by Horemuz and Jansson, 2016 (see link below). Its conclusions, from both analytical and empirical investigations, are that:

1) If you're free-stationing (measuring directions and distances) to establish a control new point under the total station, then the optimal (minimum uncertainty) location is near the centroid of the already-existing control points that you're using. That doesn't seem surprising.

2) If you're locating detail points (staking, topo), then the minimum uncertainty of a detail point will be near the centroid of the control points. Again, that doesn't seem surprising.

3) If you're locating detail points, then their uncertainties depend on the arrangement and number of the control points, but do *not* depend strongly on the setup location of the total station. This is the interesting result.

The authors' recommendations are that if you're doing detail surveying with a free-stationed total station, then surround your work area (detail points) with control points, but set up the total station anywhere that's convenient (e.g. able to see all the detail points), regardless of whether it's inside or outside the area that's surrounded by the control points.

I recall either Trimble or Leica having a white paper giving the same conclusions and recommendations.

http://kth.diva-portal.org/smash/get/diva2:1091655/FULLTEXT01.pdf

@bill-c

You may be thinking of this document from Trimble. Wikipedia calls this process of resection using both angles and distances "triangulateration"

@mark-mayer Yes indeed, that's the Trimble document I was recalling.

By the way, the graph in the Trimble document appears to have been taken from or inspired by the book "Surveying," Kahmen and Faig, 1988, figures 6.6.4 and 6.7.5 or 6.7.6. The Horemuz and Jansson paper made reference to this book. However, it looks like there's a discrepancy between the Trimble graph's curves for Station Points and those in the book's figure 6.6.4. While Trimble didn't explicitly label the three red curves, one would assume that they're for n = 2, 3, & 4 control / backsight points, similarly to the labeled blue curves. But, they look more like Kahmen's and Faig's curves in figure 6.6.4 for n = 3, 4, & 5.