How GKN got 20x faster in automating visual inspection

GKN Powder Metallurgy had already started to install vision systems to handle visual inspection in some of their production facilities. The company wanted to increase deployment speed of these systems and study how accuracy rates can be increased. Furthermore it was important to become less dependent (a) on highly qualified and hard to find engineering personnel that would program the software of vision systems and (b) on qualified end visual inspection personell. GKN Powder Metallurgy achieved that by deploying our computer vision and machine learning based software solution to existing vision systems and is considering to work with us on installment of our vision system (hardware and software) on new production lines.

The following defects have the highest priority for GKN: Scratches, holes, burrs, breakouts, bulgings, indentations, black spots, and imprints.

We started with an implementation and test phase that focused on one plant in Europe and a limited set of metal parts, indicative of the broad range of metal parts the company is manufacturing. We implemented our software at production speed (400ms/part) and compared its performance to legacy visual inspection systems like Keyence or Cognex. This allowed GKN to make the decision to leverage our software on multiple production lines in multiple locations.
GKN’s performance analysis of Covision’s technology proved that our technology allows them to be 20x faster in deploying new vision systems. The technology works at production speeds and includes the manufacturer’s CAD model of parts.

Working closely with the quality control and R&D departments of GKN was paramount to the success of the roll-out.

As GKN sees strong potential, Covision Quality’s software is scheduled to be rolled out in production facilities in Europe and the United States. Covision’s team facilitates the operations of scaling to various plants.