The shape of the material decides the camera before anything else does. Continuous web that never stops needs a line-scan camera building the image one line at a time, locked to the motion; discrete parts that index into place suit an area-scan snapshot. Choosing wrong makes the whole inspection fight the material.
| Factor | Area-scan | Line-scan |
|---|---|---|
| Material | Discrete parts, indexed | Continuous web, roll or coil |
| Image formation | Whole field in one exposure | Assembled line by line as it moves |
| Motion handling | Freeze with a short exposure/strobe | Lock capture to speed with an encoder |
| Coverage | A framed area, gaps between parts | Gap-free along the whole material |
| Speed limit | Frame rate and exposure | Line rate and encoder tracking |
| Typical use | Presence, assembly, packs | Print, coating, surface, registration |
The first question in an inspection design is not which camera resolves the defect, but whether the material stops. A discrete part that indexes into a station and pauses can be photographed in a single exposure by an area-scan camera — a snapshot of a framed field, frozen by a short exposure or a strobe. That suits components, packs and assemblies, where each part presents itself, is inspected, and moves on. The imaging is a series of stills.
Continuous material never presents itself; it flies past. Film, foil, paper, coil and printed web run at metres per second and cannot be paused or snapshotted without blur or gaps. A line-scan camera images them one line at a time, and by locking that capture to the material's motion with an encoder it assembles a single, gap-free image of the whole surface, however long the roll. The catch is that the assembled image is only correct if capture stays locked to the real speed — any slip stretches or smears it — so encoder integration and handling speed changes are part of the design in a way area-scan never demands.
Choosing the wrong one makes everything harder. An area-scan camera pointed at a fast web either blurs or leaves gaps between frames; a line-scan camera on discrete parts wastes its strength and complicates triggering. Getting this right first — matching the camera to whether the material stops — is what lets the rest of the imaging chain, the lighting and the model, do their jobs instead of fighting the material.
An area-scan camera captures a whole two-dimensional field in a single exposure, like a normal photograph. A line-scan camera captures one line of pixels at a time and assembles a full image as the material moves past it. Area-scan suits discrete parts that can be paused and snapshotted; line-scan suits continuous material that moves without stopping, because it builds a gap-free image of the whole surface line by line.
When the material is continuous and does not stop — film, foil, paper, coil, printed web — and especially when it runs fast. A snapshot of such material blurs or leaves gaps between frames, whereas a line-scan camera locked to the web's motion by an encoder builds a single, undistorted, gap-free image of the entire surface however long it is. The trade-off is that capture must stay locked to the real speed, so encoder integration is part of the design.
Because a line-scan image is assembled one line per fixed advance of the material, and each line must correspond to a known amount of travel. If capture is not locked to the actual speed by an encoder, any variation in speed stretches or smears the assembled image and distorts measurements. The encoder keeps the image geometrically correct even as the line speeds up or slows down, which a paused area-scan snapshot never has to worry about.
Written for engineers. Share it with one.
Tell us the specification, the constraint and the deadline. Programmes that cross silicon, radio, embedded and AI are where Faststream is strongest.