Help Me Understand Resection

There have been several discussions of this topic that show up in a search:

https://surveyorconnect.com/community/?wpfs=resection

I routinely use resectioning. But the only time I do it with 2 points is for searches and similar non-critical positioning tasks. When I use it for precise layout or boundary ties I resect to a minimum of 3 points. With 3 points I get residuals, which gives me a really good reading on how precise the location is.

If doing 2 points you would want the triangle formed by the 2 points and the instrument position to be?ÿ equilateral, and to confine the points to be tied to inside that triangle. With 3 points you form 3 such triangles, and a much larger area to work within.?ÿ In practice to can stretch that some.?ÿ ?ÿ?ÿ

Taboo?

?ÿ

My 'rules for resections' are the same as every other topic. It depends. I run the risk/reward computation in my head on how to get stsrted, and end with sufficient checks to prove up what I've done.

With modern instruments it is possible to do extremely tight work from resections with an angle near 180 degrees. I wouldn't try that with my t1 and top mount, and certainly not with a black transit. The absolutes of yesteryear are fading fast..?ÿ

Posted by: @thebionicman

With modern instruments

Modern instruments get you distance and angle. Decades-old discussions of resection may be considering only angles, as that was often covered in books.?ÿ

Having both kinds of measurement pretty much removes the requirements on angle between control points, but you still shouldn't go much outside the immediate area.

Did a lot of angle only resecting in the '70s and considered it only good to a few tenths when measuring locally, fine for bringing in construction control.?ÿ Then EDMs showed up which tightened it to as good as control traversing in the salient area.

BTW, consider that GNSS surveying is pure trilateration from "stations" many thousands of miles distant travelling at high speeds.?ÿ I know many will disagree but for me GNSS is only good to a few tenths using distant control and worthless when setting building control, concrete and accurate relative elevations for construction.?ÿ Steel tape (maybe EDMs with good targetry), a level and a theodolite reigns supreme when working in realms of a few hundred feet.

Feel free to beat me up.

?ÿ

I've only used resectioning on small construction projects where maybe a dirt pile is blocking my view and will be gone by the end of the day.?ÿ I think it's a great tool and should definitely be part of your bag of tricks.

In my experience Resections can work, if you have a well-calibrated instrument. We use 1" and 0.5" guns. We have found 3" and 5" instruments just don't cut it. It's not just the angle precision of your instrument. The EDM and even the optical plummet are also important.?ÿ

I use resection all the time with our Trimble SX10, mainly when performing scans. My typical MO is to set up the instrument and perform the first resection off of three RTK positions, ideally around 90 degrees or a little more apart from one another angularly. Then if possible I shoot in several remote points using reflectorless measurements that can be used for additional resections. We also use high precision traverse kits with nodal prisms and good tripods. I can normally expect the the residuals from my initial resection to be less than 0.01' in all three dimensions, and usually on the order of 0.003' in the horizontal components; further resections and/or traverse points are better than that.

I try to keep 90 degrees or close to it between the points for the same reason you should when calculating any kind of intersection; the possible area of intersection between two points is the lowest when they're 90 degrees apart.

It's really just one of those things you need to go out and practice until you're comfortable with the procedure and the results. The SX10 is an extremely precise instrument; I can do things with it that I wouldn't try with a lesser gun. And (it should go without saying) all measurements involving resection or traverse are made in both faces.

Resection. Angle only resection. Itƒ??s what gave resect a bad name. Adding in the edm changed that. And, some redundancy is good too.?ÿ
Resection is like watermelon. When done properly, can yield excellent results. When done poorly, can make you hate watermelons.?ÿ
thank you

Nate

@mike-marks I agree that GNSS isn't accurate enough for high precision building and construction layout. I don't agree at all with the "only good to a few tenths" part; GNSS is cm level when used properly, especially the "latest and greatest" receivers. This is easily provable.

I am a big fan of resections. It is not necessary to set up over a control point, as long as you can sight a minimum of two (preferably three) points with decent geometry. Put the instrument where it needs to be for the work.

Second, it is possible to eliminate centering errors and measure-up errors by entering an instrument height of zero. This is critical when using resections for tasks like monitoring surveys or precise stakeout, because in any setup over a point, those two error sources are not eliminated by repeating angle and distance measurement. Properly done, a zero-height resection is better than a single-backsight solution and will yield better results. As others have mentioned, a three-point resection is best practice for such work. We will use anywhere from three to eight control points to position for typical monitoring work.

So there's no need to set a point in the ground for a resection, unless you need it in your control network and you will be coming back to it regularly. If you do set a point and measure up, you are negating the benefits of the zero-height setup. This is something I still argue about with surveyors who cannot conceive of a setup without a point underneath the instrument.

Measuring both angles and distances instead of angles-only has greatly improved what was traditionally a mediocre solution. Most field software will give you residuals and your positional standard errors after running the resection routine.

Also, be certain to understand how your survey control computes the resection and verify the residuals. For instance, we use 2D+H settings with no computed scale factor. I suggest not letting the controller scale your results, it may make a poor solution look good enough to accept.

Posted by: @rover83

I am a big fan of resections. It is not necessary to set up over a control point, as long as you can sight a minimum of two (preferably three) points with decent geometry. Put the instrument where it needs to be for the work.

So there's no need to set a point in the ground for a resection, unless you need it in your control network and you will be coming back to it regularly. If you do set a point and measure up, you are negating the benefits of the zero-height setup. This is something I still argue about with surveyors who cannot conceive of a setup without a point underneath the instrument.

?ÿ

Agreed and the example I use is for a scan using the SX10 I just check into 3 or more control points and go.

It's important to keep pushing the narrative of know why and what you're doing for the consideration and consistency of the data collection required.

Button pusher surveying can get deeply mired without proper guidance and usage of concepts and techniques that were developed without button pushing convenience.

Posted by: @rover83

So there's no need to set a point in the ground for a resection, unless you need it in your control network and you will be coming back to it regularly. If you do set a point and measure up, you are negating the benefits of the zero-height setup.

For the last couple of years I've been doing monthly (wet months) and quarterly (dry months) settlement monitoring on a storm drain pump station.?ÿ With all the site obstructions there's only 1 spot from which I can see all of the monitoring stations with the same rod height, with only a couple of tenths of leeway in the setup location.?ÿ Rather than try to find that spot each time by estimation, on Day 1 I set a point in the ground, and I set up over it each time.?ÿ I do measure the HI with a tape, but I don't hold the original point coordinates, I assign it a new point number and let the subsequent resection adjustment determine its position.

Having a consistent height comparison for the control point gives me one more bit of reality check on each monitoring event.?ÿ If I were to use the zero-HI method I wouldn't have any basis for height comparison between occupations.

Also, on topo surveys I always include a realistic HI (measured or estimated) on resected points so I don't have to remember to exclude that control point from the terrain model.

I'm not a fan of the zero-HI approach.?ÿ I just don't see any reason not to use a realistic HI.

@jim-frame

With all the usage and even necessity with this location seems to hold, why not make a secondary control point and monument it?

Just curious, I'm still working on learning about what I don't know.

Posted by: @jim-frame

Also, on topo surveys I always include a realistic HI (measured or estimated) on resected points so I don't have to remember to exclude that control point from the terrain model.

I'm not a fan of the zero-HI approach.?ÿ I just don't see any reason not to use a realistic HI.

I use only selected points in my DTM, and all points are sorted to layers by the Descriptor Keys.?ÿ Control points are not being used (also - no boundary points, no utility points, no building shots, etc.), and it's easy to do.

I have a different code for resected (and unmonumented) instrument positions. So they sort to their own layer.?ÿ

I'm collecting a lot of things these days by reflectorless. Basically anything that sticks up. Building walls, signs, hydrants, risers and kiosks, trees, etc. No elevation on such things is valid for DTM, of course. SO the layering strategy is a necessity.?ÿ?ÿ

?ÿ

There is a big difference between resectioning and what some survey packages use such as a "two point free station". The two point free station is much better but still may not always provide the precision desired. At least once a single two point free station did not provide me with satisfaction. I redid the two point free station swapping the order of points, then meaned the two new locations. That meaned position provided me with mathematical satisfaction.

Paul in PA

Posted by: @norman-oklahoma

I'm collecting a lot of things these days by reflectorless. Basically anything that sticks up. Building walls, signs, hydrants, risers and kiosks, trees, etc. No elevation on such things is valid for DTM, of course. SO the layering strategy is a necessity.

We do the same thing, but use point groups for sorting. Add a unique character or two to a code, and it is sorted out of the "Surface Points" group. Import CSV and linework, update point groups, add breaklines to surface, rebuild surface. Done.

Posted by: @jim-frame

I'm not a fan of the zero-HI approach.?ÿ I just don't see any reason not to use a realistic HI.

For the sake of precision, if you are letting the point float every time you monitor, why introduce additional error into the adjustment? Why not eliminate the centering and measure-up errors?

For most recent monitoring projects I have done, we would not have been able to detect maximum acceptable movement without using zero-height resections. No matter how many points we tie or sets measured, centering and measure-up errors are going to stick around for the duration of the setup.

And monitoring is often a separate task for a larger construction project. In that case my monitoring control might not even be in the same coordinate system as the rest of the project - often I orient it to whatever we are monitoring for ease of movement detection.

Sure, sometimes we'll set a nail at the ideal instrument placement. If we can use that point for later work, I'll have the crew set up over it and record HI in the field books, but always enter zero in the instrument height. If I decide to bring that point (and the other monitoring control points) into the larger control network system, I will process several monitoring sessions with HIs separately against the project control.