16 November 2026

What GSD do you need to read hairline cracks in masonry?

Close-up of a lichen-covered stone wall with a mullioned Gothic window, showing fine cracks between the blocks.

If you've been asked to justify a survey spec to a committee, you've probably run the numbers yourself: a hairline crack in masonry can be under a millimeter wide. A drone flying at 10 cm GSD puts a hundred millimeters of real-world ground into a single pixel. On paper, that gap looks disqualifying. In practice, it's the wrong question to be asking about a condition baseline.

The resolution math, and why it doesn't settle the matter

To literally resolve a 1 mm crack as a distinct line in an orthomosaic, you'd need a GSD well below the crack width, call it sub-millimeter, which means a terrestrial rig or a very low, very slow pass with a long lens, flown close enough that you're really doing elevation photography, not a site survey. That's a different job with a different cost, and most authorities don't have the budget or the access to run it across every elevation of a scheduled monument every two or three years.

Sub-10cm GSD from drone or crewed aircraft RGB imagery isn't aimed at reading a crack in a single frame. It builds a photogrammetric model you can measure from, consistently, survey after survey. That distinction matters more than the pixel count. A hairline crack rarely stays hairline. What a conservation officer needs to catch is the crack opening, a lintel deflecting, a parapet leaning, a cornice losing material, stone spalling at a joint. A measurable 3D model captures each of these as a change in geometry, and geometry is something you can pull a number from, rather than eyeball off a photo.

What a baseline-plus-review model catches

Run a survey now and you get a baseline: a dimensionally accurate model of the monument as it stands, at a known GSD, that you can pull measurements from directly rather than guessing off a photo. Run the same survey again at the next condition review, flown to the same resolution, and you can compare the two models point by point. A crack that's widened by a few millimeters over two years shows up as a measurable displacement between the two surveys, even though neither individual image resolves the crack line itself. A bulge in a wall face, a few centimeters of settlement at a buttress, the slow tilt of a chimney stack: all of that is visible as change between baseline and review, which is the thing a condition record needs to show an inspector or a grant panel.

This is also why the "informal photo comparison" approach that most authorities fall back on doesn't hold up. Standing in the same spot with a camera and eyeballing this year's photo against last year's catches gross, obvious change and misses almost everything else, because the angle, the light, and the lens never match exactly twice. A measurable model removes that guesswork. You're comparing coordinates, not memories of where you stood.

None of this means fine cracking goes undocumented. Where a crack is already known and flagged, close-range photography or a monitoring gauge still does that specific job, and a drone survey isn't trying to replace it. What the survey gives you is the record for everything else on the structure, elevation by elevation, at a resolution fine enough to catch movement before it's an obvious problem, delivered once as a baseline and again at each review so the comparison is built in rather than improvised.

If you're specifying a survey and trying to decide what GSD actually earns its keep for your monument, Heritage Documentation builds exactly that baseline-and-review pair from drone or aircraft RGB imagery. Get in touch if you want to talk through what resolution your structure needs before your next condition review is due.

Get started

← Back to the blog