Walk past any under-construction building, and you’ll probably notice the scaffolding, the cranes, maybe the workers moving around the upper floors. What you almost never notice is the thin green mesh hanging off the edges of each level. It’s called a debris net, sometimes a safety net or a “Feri-Feri net” depending on which site you’re standing at, and its whole job is : catch whatever falls off the building before it hits the ground, a vehicle, or a person walking by.

Most of us walk right under these nets without a second thought. But if you’re the site’s engineer responsible for that stretch of footpath, that net is one of the few things standing between “normal day at work” and “incident report.” And here’s the part that surprised me when I first started looking into this — these nets are almost never checked properly. Someone glances up from the ground, decides it “looks fine,” and moves on. Nobody’s climbing out to the edge of a fourth floor slabs every other day just to tug on a net.

That gap — between how important these nets are and how little attention they get — is exactly what pushed us toward using drones for this.

Why a Net That “Looks Fine” From Below Often Isn’t

Debris nets take a beating. They’re outdoors, tied to scaffolding pipes or rebar, exposed to sun, rain, wind, and whatever concrete dust and construction debris lands on them daily. On a lot of sites, the standard practice is to clean them out every two weeks — clear off the collected rubble, check the ties, re-tension if needed. That’s the theory anyway.

In practice, two weeks can stretch into three or four if the site is busy, if the cleaning crew is short-staffed, or if nobody’s flagged that a particular net needs attention. And the net doesn’t fail at all at once. It sags a little first. One or two of the tie points loosen; the fabric starts drooping under the weight of collected debris, and from ground level it still looks like “a net is there,” so it doesn’t raise any alarms. By the time someone notices it’s sagging badly, or worse, that a section has come completely loose, it’s already been a risk for days.

This is the exact problem we were trying to solve when we started flying drones along the building perimeter with an automated net-checking system built into the footage.

What the Drone Actually Looks For

The idea is fairly simple once you see it in action. A drone flies along each floor level of the building, capturing the net at close range, and the analytics system processes that footage in real time, overlaying information directly onto the video. You can see this in pictures below — that’s the actual output from one of our site flyovers.

Every frame gets scored on a few things:

• **Net Status** — whether a net is present in that section at all
• **Visible Coverage** — how much of the expected net area is actually visible and intact
• **Sagging Score** — a number that reflects how much the net is drooping compared to a properly tensioned net
• **Gap Score** — whether there are holes or open sections in the mesh
• **Risk Level** — OK, WARNING, or missing, based on the combination of the above

In the first image, you can see the sagging score sitting at 0.145 with visible coverage around 22%, and the system has already flagged it as a WARNING with the reason listed plainly: “Possible loose/sagging net shape.” That’s not a guess — you can see the dip in the net outline where it’s pulling away from the horizontal line it should be holding.

The second capture is from a different stretch of the same building, and it’s worse. Sagging score has climbed to 0.213, coverage is close to 29%, and again it’s marked WARNING for the same reason. If you look closely at the net’s edge outline traced in green, you can see exactly where it dips — that low point in the middle is where the ties have loosened enough that the whole net is sinking under its own weight and whatever debris it’s been collecting.

Now compare that to this one:

Sagging score here is 0.008 — practically flat. Coverage is around 22%, similar to the first image, but the risk status reads OK, and the reason simply says “Net detected.” This is what a properly tied net looks like from the air: a clean, mostly straight line along the top edge, no major dips, no loose sections pulling it out of shape.

Side by side, the difference is obvious even to someone with no engineering background. That’s kind of the point — the system is designed so that a site supervisor doesn’t need to interpret complicated data. Green means fine. Yellow warning means someone needs to go check that section before the next scheduled cleaning. And if a stretch of the building has no net at all, it gets flagged separately as a “no net” zone, which is treated as the highest priority because that’s an open fall zone with nothing catching debris underneath it.

Sagging vs. Missing — Two Very Different Problems

It’s worth pulling these apart because they get treated differently on site, and they should.

A sagging net is still doing part of its job. It’s there, its catching things, but it’s not tensioned the way it should be. Left alone, a sagging net usually gets worse — the sag increases, more weight accumulates in the low point, and eventually a tie point can give way entirely. This is a “fix it soon” issue, not a “stop work” issue, but it shouldn’t be ignored for another cleaning cycle.

A missing net, or what our system logs as “no net,” is a different category altogether. That’s an area where debris has an unobstructed path from the working floor down to the ground. Maybe the net was taken down for some access reason and never put back up. Maybe it tore away entirely, and nobody noticed because, again, nobody’s checking these things closely from the ground. Either way, this needs an immediate response, not a scheduled one.

Having drone footage that separates these two categories automatically means the site team isn’t treating every net issue the same way. A sagging net at 0.145 gets a note and a fix-it-this-week priority. A missing section gets flagged immediately, and someone walking out there that same day.

Why the Two-Week Cleaning Cycle Isn’t Enough on Its Own

The two-week cleaning schedule makes sense from a labor and logistics standpoint — you can’t have someone out there every single day physically inspecting nets on every floor of a multi-story building. But two weeks is a long gap when you consider how fast a sagging net can go from “slightly loose” to “actual risk.” Wind events, heavy rain, or just the steady accumulation of debris weight can shift a net’s condition well before the next scheduled cleaning rolls around.

This is really the gap drone analytics is filling. It’s not replacing the physical cleaning crew — someone still must go out there, clear the debris, re-tie the net, and do the hands-on maintenance. What the drone flyovers add is visibility in between those cleaning cycles. Instead of waiting two weeks to find out a net has been sagging badly for the last five days, the site team gets a flag much sooner, with an actual coverage percentage and sagging score attached to it instead of someone’s guess from ground level.

What This Looks Like on an Actual Site

In practice, the workflow is straightforward. The drone does a pass along each floor’s perimeter — this can be scheduled as often as makes sense for the site, whether that’s every few days or tied to specific weather events like a storm that just passed through. The footage gets processed, each net section gets a status, and anything flagged WARNING or missing gets routed to whoever’s managing site safety that week.

From there, it’s a normal maintenance task. Someone climbs out, checks the flagged section, re-ties or re-tensions as needed, and the net goes back to OK status on the next pass. The value isn’t really in the technology being flashy — it’s in making sure a loosely tied, sagging net doesn’t sit there unnoticed for a week and a half because nobody happened to look up at the right angle from the ground.

The Bigger Picture

None of this is meant to replace the judgment of the people actually running the site. A drone flyover doesn’t know if a piece of scaffolding pipe shifted for a legitimate reason, and it won’t call in a repair crew on its own. What it does is remove the guesswork around “is that net actually fine, or does it just look fine from down here.” Given how easy it is for a sagging net to go unnoticed between cleaning cycles, having a consistent, repeatable way to check every section — with an actual number attached instead of a shrug — makes it a lot harder for a real risk to slip through unnoticed.

Debris nets aren’t the most exciting part of a construction site. Nobody’s writing case studies about how well a net holds up. But they’re doing quiet, important work every single day, and checking on them properly, rather than assuming they’re fine, is one of those small operational habits that ends up mattering a lot more than it seems like it should.