What an orthomosaic is — and why it isn't just a big photo

Take one photo of a job site from a drone and look closely. The middle looks right. Toward the edges, light poles and building corners lean outward, a berm looks stretched on one side and squashed on the other, and the scale is different in each corner. That's perspective and terrain doing what they always do to a camera — the same reason a wide-angle selfie distorts a face. It's a useful picture. It is useless for measuring.

From a photo to an orthophoto

An orthophoto is an aerial image that has been geometrically corrected so that every pixel is placed where it would be on a map — as if the camera had been directly overhead of every single point at once. Propeller Aero's explainer on the subject puts it simply: the correction removes the distortions that perspective, terrain, and camera angle introduce, which is what turns a picture into something you can measure on. Once that's done, the distance between two pad corners on screen is a real distance on the ground.

Stitch hundreds of corrected photos together and you have an orthomosaic: one seamless, measurable image of the entire site. In day-to-day use the words blur together — most people just say "the ortho".

How it's made

• Capture. The drone flies a pre-planned grid and takes photos with heavy overlap (commonly 70–80%), so every spot on the ground appears in many frames from slightly different angles.
• Photogrammetry. Software finds the same features across overlapping frames and triangulates both the camera positions and the shape of the ground beneath them.
• Georeferencing. Ground control points (surveyed targets visible in the photos) or corrected GPS on the drone (RTK/PPK) pin the model to real-world coordinates.
• Orthorectification and mosaicking. Each photo is reprojected onto the terrain model and blended with its neighbours into a single image, with seams placed where they won't cut through anything important.

What the numbers on a deliverable mean

GSD — ground sampling distance — is the size of one pixel on the ground. Propeller cites 1–3 cm as typical for construction work; our deliverables usually land around 0.8 in/px (about 2 cm). A smaller GSD means more detail, and it is the figure most people fixate on. It is not accuracy. Propeller makes the same point: a crisp map can be sitting several feet from where it belongs if nothing tied it to the ground. Accuracy comes from ground control, GPS corrections, sensible flight planning, QA during processing, and doing all of it the same way every flight.

Coordinate systems trip up more projects than resolution ever will. Most drone workflows output WGS84, the global system GPS uses. Your civil drawings are probably in a state plane or local site system. If the two aren't reconciled, your design overlay will look shifted even though both files are individually correct. Confirm the coordinate system with your engineer or surveyor before the first flight, not after the first argument.

What project teams actually use it for

• Progress documentation — the same site, the same grid, week after week, so "what was there on the 14th" is a click, not a debate.
• Design overlays — drop the site plan on the ortho and see whether the pad, utilities, and laydown areas are where the drawings say they are.
• Quantities — paired with the surface model, the ortho is where you outline a stockpile or a cut area to get a volume.
• Records — timestamped, georeferenced images are hard to argue with in a delay claim or a dispute with a sub.
• Communication — owners and lenders understand a map instantly; they do not understand a folder of 600 photos.