How Can Vision Systems Improve Quality Control in Manufacturing?

Inspection Needs Consistency

Quality inspection can be quite a challenge when dealing with fast-moving products, small defects or repetitive checks that must be made thousands of times. While humans are an invaluable asset for judgment and perception, visual fatigue and the speed of production can combine to make a very challenging inspection task. Alternatively, machine vision can be employed to perform a specific visual inspection test over and over again.

Cameras Become Measurement Tools

An industrial vision system employs cameras and other optics to capture images for analysis. Depending on the job, an inspection system can identify missing components, check the orientation of a part, read barcodes, measure size, count items or even identify defects in appearance. The system is usually programmed with the information needed to evaluate the item being checked, rather than simply recording a video stream.

Lighting Is Often the Hidden Key

A high-quality camera is of little use if the image isn’t captured properly. Reflections, shadows, varying ambient light and low contrast between an item and its background can all make a defect difficult to see. A properly designed vision-system inspection should carefully consider lighting and positioning so that the feature being inspected is visible and reliable from cycle to cycle.

Vision Can Guide Robots Too

Vision is not limited to pass-or-fail inspection. It can also tell a robot where a part is located or how it is oriented. AKA Electrical & Automation has demonstrated integrated robotic vision applications involving part identification, placement and assembly, as well as uses such as defect detection, measurement, barcode scanning and parts counting.

False Rejects Need Attention

An inspection system which frequently rejects good products is potentially wasteful and frustrating for operators. A system which is too lax can miss defects. Testing should be done with normal variations of colour, position, finish and presentation to ensure thresholds are appropriate.

Data Can Improve the Process

Vision systems can record the types and frequency of defects, providing insight into where process instability has occurred. Should an increase in one defect type occur following a change in tooling or materials, the vision system can help identify this discrepancy more quickly and precisely.

The Goal Is Better Process Control

Machine vision is most valuable when it contributes to more than simply removing manual inspection. The ability to provide consistent and repeatable checks, traceability of information and immediate results can give manufacturers insight into their processes. The most effective systems are built from a consideration of a particular quality question and fit into the production line to provide actionable data to the right people.