Purchase orders accepted from universities, colleges and government agencies. Financing available with Affirm. W-9 and vendor pack
×
MySurveyingDirect
+1 888 850 6533Mon to Fri, 9 to 5 MT 0
Home/Resources

Resources

Short, practical notes on the things that actually decide whether a survey holds up. No product pitches. If something here saves you from buying the wrong instrument, that is a good outcome for us too.

Accuracy, explained honestly

Accuracy is the most abused number in this industry. Four different things get called "accuracy" on a specification sheet and they are not comparable.

Term What it actually means Why it gets misused
Relative accuracy How well the cloud agrees with itself. Measure a wall in one scan, measure it again, how close are they. It is always the flattering number. A scanner can be excellent relatively and be two meters off the national grid.
Absolute accuracy How close a point is to its true position in a real coordinate system. Rarely quoted, because it depends on your control and your setup, not on the instrument alone.
Precision or repeatability How tightly repeated measurements cluster. Says nothing about whether they cluster in the right place. Quoted as accuracy constantly. A consistently wrong instrument is precise.
Nominal or laboratory figure What the instrument achieves under ideal conditions, usually short range, good geometry, open sky. Your site is not a laboratory. Expect field performance to sit outside the quoted figure.

The questions worth asking about any accuracy claim

  • At what range? A LiDAR accuracy figure without a distance attached is meaningless. Accuracy degrades with range on every instrument ever made.
  • At what reflectivity? Manufacturers quote range at 10% reflectivity, or 50%, or 80%. The same sensor can honestly be described as reaching 250 meters or 450 meters depending which one is chosen. Check before comparing two sheets.
  • Prism or non-prism? On total stations these are wildly different numbers. A 4,000 meter prism range and a 600 meter non-prism range describe the same instrument.
  • One sigma or two? A figure quoted at one standard deviation covers about 68% of measurements. The same instrument quoted at two sigma looks twice as bad and is twice as honest.
  • Under what satellite geometry? A GNSS precision figure assumes open sky. Under canopy or between buildings, all bets change.
Our own rule for this website: where a manufacturer publishes a figure, we quote it and name the condition it was measured under. Where they do not publish one, we say "not published" rather than estimating. If you find a number on this site you cannot trace back to a manufacturer specification, tell us and we will correct it.

RTK, PPK and static corrections

A GNSS receiver on its own is good to a few meters. Everything better than that comes from correcting against a second known position. There are three ways to do it, covered side by side on the compare page. The practical points:

  • Log raw data even when working in RTK. It costs nothing and turns a dropped correction link from a lost day into a processing job.
  • Baseline length matters. The further you are from the base, the more the atmosphere between you and it degrades the solution. Emlid publish centimeter accuracy to 60 km in RTK and 100 km in PPK on the RS2+, which gives you a sense of the shape of the curve.
  • A network subscription is not always better than your own base. On a site with poor cell coverage, or one you will return to repeatedly, your own base on a known point is faster and more reliable.
  • Convergence time is a real cost. Under five seconds sounds like a detail until you are measuring several hundred points in a day.

Setting up your first base station

The single most common cause of a survey being wrong is not the instrument. It is the base setup. In order:

  1. Choose the point deliberately. Open sky, away from buildings, fences and vehicles that reflect signal. If you can see a lot of sky from the antenna, you are most of the way there.
  2. Decide whether the point is known or not. If you are setting up on a known control point, enter its coordinates. If you are not, the base will average its own position, and everything you measure will be internally consistent but sitting on an arbitrary origin. Both are valid. Confusing the two is not.
  3. Measure the antenna height, then measure it again. Write it down. Note whether the figure you are entering is to the antenna reference point or to the ground mark, because the software makes an assumption and it is not always the one you made.
  4. Let it settle before you start. An averaged base position gets better with occupation time. Walking off after thirty seconds is where a lot of vertical error comes from.
  5. Check into a second known point before you trust anything. One independent check takes two minutes and catches almost every setup error there is.
  6. Record what you did. Base point, antenna height, coordinate system, correction source, time. When something looks wrong three weeks later, this note is the difference between a five minute answer and a resurvey.
The vertical is where it goes wrong. Horizontal errors tend to be obvious. Vertical errors from a mis-measured antenna height are exactly the size of the mistake, they are consistent across the whole job, and nothing on screen looks unusual. Measure it twice.

More

Something here not covered, or not clear?

Call +1 888 850 6533, Monday to Friday, 9 to 5 Mountain Time, or email support@mysurveyingdirect.com. You will get a specialist, not a script.