How Often Should You Re-Fly a Site? Setting a Drone Monitoring Cadence

Published 2026-09-30 · UAV Imaging Inc.

Key takeaways

Cadence follows the decision the data feeds. If a number is reported monthly, a quarterly flight means two of every three reports run on an estimate.
Consistency beats frequency. Four flights a year on the same plan, the same reference and the same base surface tell you more than eight flights that cannot be compared.
Mobilization is the fixed cost of a drone visit, not the flying, so a standing schedule and grouped sites cost far less per visit than booking each one when it becomes urgent.
Build event-driven flights into the schedule rather than treating them as exceptions. A dispute, a major haul, a heavy rain or a shutdown are all worth a capture regardless of where the calendar sits.
Pick a cadence, run it for three or four cycles, then change it once with a reason. Most sites are re-flying either far too seldom to inform anything or often enough that nobody reads the output.

Once a site has been flown a second time, the question stops being whether drone data is useful and becomes how often to collect it. It usually gets answered badly, in one of two ways. Either a number is set because it sounds responsible, most often quarterly, or flights are booked reactively when someone upstairs asks a question that nobody can answer.

Neither is a schedule. Here is how to arrive at one.

Cadence follows the decision, not the calendar

Start from the other end. What decision does this data feed, and how often is that decision made?

If your inventory position is reported monthly, a quarterly flight means two reports in every three are running on an estimate, and the quarterly capture is not a measurement so much as a correction. If a progress draw is submitted monthly, imagery captured every six weeks will never line up with a billing period and someone will spend the difference in meetings. If a site is inspected for compliance once a year, flying it monthly produces a folder nobody opens.

The reporting cycle you already run is the best first answer to the flight cadence question, and it costs nothing to work out. Everything below either confirms it or moves it.

The four things that actually set the interval

1. How fast the site changes

An active earthworks phase and a closed cell of the same site do not deserve the same schedule. The useful test is whether the site would look materially different to someone standing on it. If two consecutive captures are nearly identical, the interval is too short for what the site is doing. If every capture is a surprise, it is too long.

2. What a stale number costs

This is the one that decides most cases and the one nobody calculates. The question is not what the flight costs. It is what you lose by carrying a number that is out of date: a stockpile position that turns out to be wrong at year end, a progress claim that gets disputed, a haul that was planned against the wrong volume, a variance that has to be explained to an auditor. If the cost of being wrong is trivial, fly seldom. If it is not, the flight is rarely the expensive item in that comparison.

3. What the data has to line up with

If the output feeds an external cycle, an audit, a year end count, a regulator submission, a lender's draw schedule, then the cadence is largely set for you and your job is to place the flights so that each capture lands before the deadline rather than after it. Timing a capture to the same cutoff as the number it will be compared against removes an entire class of argument, as our piece on reconciling drone volumes against the scale house goes through.

4. Whether anyone is going to act on it

The least discussed constraint. A capture that produces a deliverable nobody opens is not monitoring. If your organisation can absorb and act on a result once a quarter, monthly flights generate backlog rather than information. Match the cadence to the capacity to respond, then increase it when the response tightens up.

What has to stay identical between flights

This is where repeat programmes are won or lost, and it is more important than the interval. The value of the second flight is entirely in its comparability to the first.

The same flight plan. Same boundary, same altitude, same overlap, same camera settings. Reuse the plan rather than redrawing it.
The same coordinate system and vertical datum, every time. A capture delivered on a different reference is not a comparison, it is a second dataset, and reconciling them afterwards costs more than the flight did.
The same positioning approach and the same control. If ground control was used, reuse the same points. Our article on how many ground control points you need covers what that involves.
The same base surface for volumes. A stockpile volume against a re-picked base is not comparable to last quarter's. Fix the base surface once, record how it was defined, and reuse it until there is a documented reason to change it.
The same processing settings and the same reporting format. So a change in the number is a change on the ground rather than a change in the software.

Four flights a year that meet all of that are worth more than eight that do not. If you take one thing from this article, take that one.

Alberta puts real constraints on the schedule

A cadence set on paper meets a province with a strong opinion about it.

Snow cover. Surfaces under snow can be captured, but what is being measured changes and the result needs to be interpreted accordingly. Our piece on surveying stockpiles and sites under snow covers what survives and what does not.
Cold. Winter work has a defined operating envelope rather than a hard stop, described in winter drone operations in Alberta.
Smoke season. Summer smoke affects visibility, image quality and thermal work, and it does not respect a schedule. Flying surveys in wildfire smoke goes through it.
Spring melt and freeze thaw. Surfaces move. A capture either side of breakup is often more informative than one in the middle of it.
The growing season. Anything about vegetation extent has to be flown in the same part of the season each year or the comparison is measuring the calendar.

The practical response is to set the cadence and then place each flight inside a window rather than on a date, with an agreed rule for what happens when the window closes.

The economics of a repeat visit

The cost that dominates a drone visit is getting the crew and equipment to the site, not the time in the air. Three consequences follow, and all three favour a schedule over ad hoc booking.

Sites in the same area flown on the same trip cost far less per site than the same sites booked separately as each becomes urgent.
A standing schedule is plannable, which means it can be grouped, routed and fitted around other work. A reactive request cannot.
Adding a second area of interest to a visit that is already happening is cheap. If there is something you would like captured but have never justified on its own, attach it to a scheduled flight.

This is also the argument for deciding cadence at the portfolio level rather than site by site. The site that only needs one flight a year may as well be flown on the day the crew is already an hour away.

Flights the calendar will not tell you about

Build these in rather than treating them as interruptions:

Before and after a major haul, a large placement or a significant earthworks phase.
Before a shutdown or turnaround, and again after, while access is still easy.
After a significant weather event, particularly heavy rain on a site with new grading.
When a dispute starts, or when you first suspect one might. A capture from the week the question arose is worth considerably more than one from the week it escalated.
Before anything is buried, covered or built over. Underground services, subgrade, liners. That is a one chance capture.

How to start

Pick the interval your existing reporting cycle suggests. Fix the flight plan, the coordinate reference and the base surface on the first visit and write them down. Run three or four cycles. Then ask one question: in those cycles, did a result ever change a decision? If it did, hold the cadence or tighten it. If nothing changed and nothing was ever a surprise, stretch the interval and put the saving into flying more of the portfolio instead.

That review, done once, is worth more than any general rule about how often to fly a site.

Frequently Asked Questions

How often should we fly our stockpiles?

Start from how often the inventory position is reported and used. If it is reported monthly and carried into financial reporting, a monthly flight means the number is measured rather than estimated. If it is reviewed quarterly and the year end count is the one that matters, quarterly with a tighter capture around year end usually fits. The wrong answer is a cadence that leaves your most important reporting date running on the oldest data you have.

Is monthly too often for a construction site?

It depends on what the imagery is for. If it supports a monthly progress draw, monthly is the natural match because each capture lands against a billing period. On a fast earthworks phase where the site changes substantially in two weeks, more frequent capture can be justified for the duration of that phase and then relaxed. The thing to avoid is a fixed frequency held through phases that change at completely different rates.

What makes two flights comparable?

Five things, and they matter more than the interval. The same flight plan including boundary, altitude and overlap. The same coordinate system and vertical datum. The same positioning approach and the same ground control points if control is used. The same base surface for any volume. The same processing settings and report format. Change any of these and a difference in the output may be the method rather than the site, which is the failure that quietly makes a monitoring programme worthless.

Does it cost less to book flights on a schedule?

Per visit, yes, and the reason is that mobilization rather than flight time is the cost that dominates a drone visit. A standing schedule can be routed, grouped with other sites in the same area and fitted around other work. Reactive bookings cannot be planned around, arrive at short notice and often carry a trip of their own. If you have several sites in one region, setting the cadence at the portfolio level rather than site by site is usually the single biggest saving available.

Should we fly in winter or wait for spring?

Both, for different reasons. Winter work has a defined operating envelope rather than a hard stop, and skipping the whole season leaves a gap of several months in the record on a site that keeps changing. Snow does change what is being measured on a surface, so the result needs to be read accordingly rather than dropped into a trend as though it were a summer capture. The usual approach is to keep the cadence through winter, note the conditions on each capture, and keep anything vegetation related to the same window of the growing season.

How do we know if we are flying too often?

Two signals. Consecutive captures that are near identical, which means the interval is shorter than the site's rate of change. And deliverables nobody opens, which means the cadence has outrun the capacity to act on it. If neither is happening and results occasionally change a decision, the cadence is about right. Review it after three or four cycles rather than adjusting it every time a result is inconvenient.

Book a commercial drone flightMapping, inspection, multispectral and thermal across Alberta.Request a quote →or call 587-532-9000