Guy wire vs. anchor: telling them apart in aerial imagery

A guy wire and a guy anchor are two different pieces of hardware doing two different jobs, and from the air you almost never see both of them the same way. The wire is the tension member, a thin strand running from a pole attachment point down to the ground. The anchor is what's holding the other end in the dirt: a rod, an eye, a helical plate or screw anchor buried a few feet down. At sub-10cm GSD you can usually pick out the wire. The anchor itself, the part doing the actual load-bearing work underground, is invisible from a drone or a crewed flyover no matter how clean the imagery is.
So when someone asks whether a feature in an orthomosaic is "the guy wire or the anchor," the honest answer is that you're looking at the wire and inferring the anchor from where that wire meets grade. Code it as a wire ground-termination point, not a confirmed anchor, and leave anchor type and condition as something the field crew checks while they're already standing there.
What you're actually seeing at the ground end
At the point where a down guy reaches the ground, there are a few things that might show up in imagery and a few that won't.
Visible, usually: the wire itself as a bright, thin diagonal line against grass, gravel, or pavement. A guy guard or guy marker, the yellow or orange sleeve some utilities require on wires near roads or walkways, which shows up as a distinct colored segment partway up the strand. Sometimes a visible ground rod stub or turnbuckle hardware if the sun angle and resolution cooperate.
Not visible: the anchor rod below grade, the anchor plate or helix, soil disturbance from the original installation (unless it's recent). None of that is something imagery resolves, regardless of GSD. The temptation is to assume that if the wire is clear, the anchor must be too. It isn't, and no flight cadence changes that.
What this means practically: a feature class tagged "anchor" from imagery review should really be understood as "guy wire ground termination, anchor location inferred." If your register has a separate ANCHOR_TYPE field expecting rod, screw, or plate values, aerial imagery alone won't populate it. That's a field-verified attribute, not a photo-interpreted one.
Down guys vs. span guys vs. push braces
Before you even get to the anchor question, there's a step that trips up a lot of first-pass review: confirming that the diagonal line you're looking at is a down guy and not something else.
A down guy runs from a point on the pole, often near the crossarm, straight down to an anchor near the pole base, typically at an angle that puts the anchor somewhere between 10 and 30 feet out depending on pole height and loading. A span guy or stub guy instead runs between two poles or structures, roughly horizontal, and won't terminate at grade at all. A push brace (sometimes called a stiff arm) is rigid, not a wire, and angles away from the pole on the compression side rather than the tension side. Mixing these up in attribute coding means a pole gets flagged with a guy count that doesn't match reality.
Where attribute coding breaks down
Most registers carry some combination of guy count, guy direction, and anchor type per pole. In practice, those fields get populated once at construction and rarely touched again. A down guy gets added during a later upgrade, a push brace gets swapped for a guy during storm repair, and the register keeps whatever was entered years ago. From the ground, a crew would catch that in minutes. From a desk, the only way to catch it is against current imagery.
That's the comparison Grid Asset Inventory is built to run: detections from a recent flight set against whatever your register currently claims, so the mismatches surface as a list instead of a surprise during a field audit.
If you're staring at a batch of pole photos trying to sort guy wires from anchors from push braces by eye, there's a faster way to find out where your register and your imagery disagree.