Knowledge · Cameras
The first fork in camera selection is geometry: does the scene move continuously past the sensor, or do you capture a discrete field of view per cycle? This guide helps engineers choose before SKU shopping.
Decision frame
Area scan captures a full 2D frame in one exposure. Line scan builds an image line by line as material or parts move. The application motion model usually decides the geometry before you compare megapixels or interface speed.
Film, textile, paper, steel strip, conveyor surfaces: the object keeps moving. You need consistent resolution along travel direction and often unlimited length in the cross-web direction.
Typical fit: line scan camera, telecentric or large-format lens, strobe or continuous lighting aligned to scan direction.
Parts stop in a nest, a robot presents a view, or a gantry positions the camera. You capture one complete image per cycle and inspect the full scene at once.
Typical fit: area scan camera (including smart cameras and many 3D sensors), fixed lens, triggered exposure per part.
Synchronisation
Line scan only works when pixel pitch in the travel direction matches real-world scale. An encoder on the roll, belt, or linear axis tells the camera when to grab the next line so geometry stays square even if speed varies.
See also triggering and timing once geometry is fixed.
When which wins
| Situation | Line scan | Area scan |
|---|---|---|
| Unlimited length web or strip | Strong fit | FOV limit, tiling overhead |
| Stopped part in fixture | Overkill | Natural fit |
| High-res cross-web, constant speed | Efficient pixels | Needs very large sensor |
| Robot vision, discrete poses | Rare | Standard (2D or 3D area) |
| Variable line speed | Needs encoder | Trigger + strobe often enough |
Read the definition spoke for your geometry, or open the camera selection hub for optics, lighting and model shortlist.