Stakeout Basics: Navigating to Design Points with Cut and Fill

Most GNSS field work is collection: you walk to something that exists and record where it is. Stakeout is the reverse. A design gives you coordinates for something that does not exist yet (a corner, a curb line, a pipe invert), and your job is to find that exact spot on the ground and mark it. It is the workflow that turns a plan set into a construction site, and it rewards a little rigor.

The stakeout workflow, end to end

1. Import the design. Points usually arrive as a CSV (commonly PNEZD: point, northing, easting, elevation, description) and linework as DXF. Confirm the coordinate basis matches your project before anything else; a design in the wrong reference frame will stake out beautifully in the wrong place. If your corrections come from a different frame than the plan, you can inherit a 1–2 meter shift without a single error message; see NAD83 vs ITRF.
2. Select a point and navigate. The software continuously compares your rover position to the target and steers you in.
3. Mark it. When you are within tolerance, plumb the pole, set the stake or mark, and label it.
4. Record the as-staked position. Take a shot on the mark you just set. This stored point, not the design point, is your proof of what actually went in the ground.

Reading the stakeout screen

Every stakeout interface says the same few things in slightly different clothes:

• Distance and bearing to target. Far away, you get "go 41.2 m at N 63° E" or arrow guidance. Close in, this switches to fine axes such as forward/back and left/right, in feet or meters and decimals.
• Heading-up vs north-up. North-up keeps the map fixed with north at the top, which is easiest when you can relate the site to the plan. Heading-up rotates the map to your direction of travel, so "target at the top of the screen" means "walk straight ahead." Most people navigate faster heading-up and orient better north-up; learn both and switch deliberately.
• Tolerance rings. Concentric circles around the target show your acceptance radius. When your position marker sits inside the innermost ring, you are within tolerance and can mark. The ring is a decision boundary, not a decoration: set it to match the job, not the default.

Cut and fill

When the design point carries an elevation, the screen also shows the vertical delta between the ground under your pole tip and the design grade:

• Cut means the existing ground is above design grade; material must come out. A stake marked "C 0.35" tells the crew to excavate 0.35 below the mark's reference.
• Fill means the ground is below grade; material must go in. "F 0.20" means build up 0.20.

Cut/fill is only as good as your vertical chain: the design elevation's datum, your geoid model, and your antenna height all sit between the satellite and the number on the stake. Elevation confusion between ellipsoid and orthometric heights is a classic source of grief; our datum and geoid guide covers the difference.

As-staked records: your QA trail

The as-staked shot is the unglamorous half of stakeout that saves careers. For every stake, store the measured position alongside the design point it was staking, and let the software compute the deviation. The resulting report (design vs as-staked, horizontal and vertical deltas, per point) is your evidence that the layout was within tolerance on the day you left the site. When a foundation is formed in the wrong spot three weeks later, that report is the difference between "the layout was correct" and an argument. Store fix type, RMS, and correction age with each as-staked shot so the record shows not just where you staked but how trustworthy the measurement was.

Accuracy at the stake: float is not fixed

Stakeout inherits its accuracy from your RTK solution at the moment you mark, and the difference between solution states is not academic:

A float solution can look seductively steady while sitting a decimeter off. If the receiver drops to float mid-stake, stop, let it re-fix, and re-check the mark. If it will not fix, the cause is usually sky view, correction dropouts, or baseline length; our primers on RTK GNSS surveying and free NTRIP casters and CORS networks cover the diagnosis.

Field tips that prevent bad stakes

• Antenna height is the classic blunder. A wrong pole height translates directly into a wrong elevation on every stake. Verify it at the start of the session and any time the pole changes, and use software that records it with every observation.
• Stay level when it counts. A 2 m pole tilted a couple of degrees moves the tip several centimeters. Watch the bubble at the moment you mark, or use a receiver with IMU tilt compensation, which holds 1–3 cm even with a tilted pole.
• Check into control before staking. Occupy a known point first. If it checks within tolerance, your whole correction chain is validated; if not, you found the problem before it was cast in concrete.
• Re-check periodically. On long sessions, revisit a control point or a previously set stake every few hours.
• Label stakes completely. Point number, offset if any, and cut/fill. A perfectly placed but ambiguous stake still gets built wrong.

Stakeout is one of the features commercial packages most often push into paid tiers. Open Survey's plan is to ship stakeout, DXF import, code libraries, and reports free and open source, with every as-staked record carrying the full quality metadata (fix type, RMS, correction age, antenna height) that makes your deviation report worth signing.