Machine Vision

Machine vision is the use of cameras, optics, lighting and image-processing software to inspect, measure, identify or locate parts automatically in industrial processes, and to guide robots and machines based on what is seen.

A machine vision system acquires images with industrial cameras or smart cameras, using lenses and lighting chosen to make the features of interest stand out. Software then applies operations such as filtering, edge detection, blob analysis, pattern matching, measurement, code reading and, increasingly, deep learning to produce a pass or fail decision or coordinates that are sent to a PLC, robot or MES.

Typical tasks are defect inspection, presence and completeness checks, dimensional gauging, barcode and character reading for traceability, and robot guidance for picking, placing and alignment. Machine vision is used across automotive, electronics, food, pharmaceutical, packaging and textile production, often inspecting every part at line speed.

Lighting and optics often matter more than algorithms, because poor contrast is hard to recover in software. Systems are validated with controlled sets of good and defective parts and against the measurement capability the process requires. Camera interfaces are standardised by GigE Vision, USB3 Vision and GenICam, and EMVA 1288 defines how camera performance is measured.

Key points

Combines cameras, optics, lighting and software for automated decisions
Used for inspection, gauging, code reading and robot guidance
Lighting and optics are often decisive for reliability
GigE Vision, USB3 Vision, GenICam and EMVA 1288 are key standards

Where AiVibe comes in

AiAmbA use cases include visual inspection on textile, automotive and electronics lines, with existing lines retrofitted with edge devices.

Explore AiVibe’s work in Robotics & Physical AI →

Machine Vision

Key points

Where AiVibe comes in

Related terms

Terms that refer to Machine Vision