/əˈkluː.ʒən/

Occlusion

noun · AR technology

What it is

In the real world, near things block your view of far things: your hand passing in front of a lamp hides the lamp. Early AR ignored this and simply painted virtual content on top of everything, so a digital object would show through your hand as if the hand were transparent. Occlusion fixes that. When the app understands how far away real surfaces are, it can let your hand — or a mug, or the edge of your desk — correctly cover the virtual layer, which reads as depth and makes the illusion hold.

For flat-on-paper AR tracing, occlusion mainly matters at the moment your drawing hand crosses the overlay: with occlusion, your hand passes over the reference the way it would over a real photo, instead of the image ghosting through your fingers. It sits alongside the other augmented reality parts — plane tracking to find the surface and homography to pin the image to it — as the finishing touch that makes an overlay feel grounded rather than pasted on the glass.

Etymology

"Occlusion" comes from the Latin occludere, "to shut up" or "close off" — the same root behind a dentist's "occlusion" for how teeth meet. In graphics and vision it names one object shutting off the view of another. Handling it convincingly is one of the classic challenges Ronald Azuma flagged in his 1997 survey of the field, and modern frameworks address it with depth information from the camera and, on some devices, dedicated depth sensors.¹

Examples in use

In a furniture app with occlusion, a virtual chair partly disappears behind your real sofa instead of hovering in front of it. In a game, a character can duck behind your real couch. In drawing, occlusion is why your hand can move naturally across an AR-traced reference without the image bleeding through it. The free web tool has no camera and no depth, so occlusion does not apply — it draws overlays flat on an uploaded image in the browser.