Vision inspection systems use cameras, lighting, optics, and image-processing software to inspect products directly on a production line. Each system captures a triggered image of every part, compares it against defined criteria, and signals the line's PLC to pass or reject, catching defects, missing components, and dimension errors at the station instead of at shipping.

Quick Facts

Core checks. Defect detection, presence/absence, dimensional measurement, code reading, and assembly verification.

Speed. In-line systems process hundreds to thousands of parts per minute; line-scan platforms run up to 600 m/min.

Resolution. Rule-based 2D handles surface checks, 3D evaluates depth and geometry, and precision platforms resolve features down to 1.5 micrometers.

Output. A pass/fail signal to the PLC plus stored images for traceability.

Hardware. Smart cameras such as Cognex In-Sight, or camera-and-controller setups such as FANUC iRVision inside robot cells.

How does a vision inspection system work?

A photo-eye or encoder pulse triggers the camera as each part enters the field of view, and the image is captured under controlled lighting chosen to make the defect visible: a backlight for silhouettes and dimensions, a dome for glossy surfaces, structured light for depth. The software, rule-based for well-defined features or AI-trained for variable ones, compares the image against the inspection criteria and sends the result to the line PLC over EtherNet/IP or discrete I/O. Failed parts are removed by an air blast, pusher, or divert before the next station, and every image is logged against the part for traceability. Lighting is the make-or-break variable; most machine vision inspection failures trace to ambient plant light washing out the feature, not to the camera.

Vision Inspection System Components

 

What is machine vision inspection used for?

Defect detection. Surface flaws, scratches, dents, and contamination at line speed.

Presence and absence. Missing fasteners, labels, seals, and components before the next station.

Dimensional measurement. Non-contact gauging of critical dimensions on every part, not a sample.

Code reading. 1D and 2D code verification for lot tracking and traceability.

Assembly verification. Confirming the right part is in the right place before value is added downstream.

Fill and packaging. Fill level, cap presence, and label position on packaging lines.

Why do vision inspection systems beat manual inspection?

A vision inspection system applies the same criteria to every part, at line rate, on every shift. Human inspectors fatigue, and catch rates drift between operators and across a shift; automated inspection holds accuracy in the micron range at hundreds of parts per minute and records every result. The economics follow the escape: a defect caught at the station costs a rejected part, while the same defect caught at shipping, or by your customer, costs rework, freight, and the next order.

What does it take to integrate vision inspection on an existing line?

Four integration decisions determine whether the system performs: where in the line the inspection runs, how the trigger synchronizes to line speed, how the lighting defeats ambient plant light, and how the pass/fail handshake reaches the PLC and the reject station. The camera is the smallest part of the project. On robotic lines, the same engineering extends to guidance, with iRVision locating parts for FANUC arms inside the work cell.

Add Inspection to Your Line

HOJ integrates vision inspection systems into conveyor and robotic lines, engineering the camera, lighting, PLC handshake, and reject station as one project, with in-house controls engineering since 1964. Talk to an automation engineer today.