What sets an industrial camera apart from an ordinary one is, above all, reliability: it captures an image exactly at the moment it receives a trigger, runs continuously, usually transmits uncompressed images and has industrial connectors. The camera is only one part of a machine vision system. The result depends on the combination of camera, lens, lighting and software.
Camera types
- Area-scan 2D cameras. The most common type: one frame captures the whole part. Suitable for inspection, reading characters, QR codes and displays, and determining position.
- Line-scan cameras. Capture one line at a time, and the image builds up as the object moves. Used for long or continuous products: strips, sheets, rolls.
- 3D cameras and scanners. Provide not just an image but also the distance to every point. The main technologies are structured light (e.g. Photoneo PhoXi and MotionCam-3D), time-of-flight, ToF (e.g. Basler blaze), laser triangulation and stereo vision.
- Specialised cameras: high-speed, thermal, hyperspectral. They are needed less often and are chosen for a specific task.
How to choose a camera
- Smallest defect or dimension. Together with the field of view, this determines the required resolution. As a rule, the smallest feature should cover at least a few pixels.
- Motion speed. This determines exposure, frame rate and whether a global shutter is needed.
- Lens. Focal length, working distance, depth of field. Telecentric lenses are often used for measurement.
- Lighting. Often more important than the camera itself: directional light, backlighting, dome lighting, polarisation.
- Interface. GigE allows long cables; USB3 suits short distances and fast image transfer.
- Environment. Dust, vibration, temperature, ambient light. A protective housing is sometimes needed.
Camera calibration
Calibration maps pixels to real millimetres and removes lens distortion. For measurement it is essential. If the camera guides a robot, hand-eye calibration is also performed: it determines where the camera is relative to the robot, so that a point calculated from the image matches robot coordinates. For 3D scanners, depth accuracy is also checked across the whole working volume.
Where industrial cameras are used
- Quality inspection: defects, completeness, labels.
- Measurement: thickness, diameter, length, position.
- Reading: QR codes and barcodes, characters, instrument displays.
- Robot guidance: finding objects in a pile and calculating the pick point.
Our examples include reading meter LCDs on a metrology bench, sorting logs by thickness, and finding logs in a pile with a 3D camera for a robot. We start vision projects with a trial on your parts, so you can see the result before any larger investment. Find out more in our machine vision services.