From target/actual comparison to AI-controlled circuit board inspection

Our customer, a leading manufacturer of PCBs in the field of high-quality power electronics, was faced with the challenge of precisely inspecting increasingly complex PCBs. Classic defects, such as traces that were too thin, interruptions or missing vias, previously required a great deal of manpower and a combination of human and machine testing. The increasing variety of types and higher quality requirements made a more efficient and precise solution necessary.

With our machine vision solution and the use of artificial intelligence, we offer the perfect solution: the aku.visionManager®. The high-resolution line scan cameras supplied by Chromasens GmbH deliver images in the smaller micrometer range per pixel, which are razor-sharp and visualize even the smallest details. Instead of the conventional target/actual comparison, we rely on neural networks. These learn how defects appear on the circuit boards and recognize anomalies independently. The system is constantly evolving and improves with each new inspection.

The use of aku.visionManager® enables the customer to significantly reduce time and personnel costs. There is no need to create templates, as the software is adaptive and can be customized to the customer’s specific requirements. The more checks are carried out, the better and more accurate the system becomes. This leads to a significant increase in inspection accuracy and efficiency. The combination of the customer’s many years of PCB expertise and aku’s machine vision know-how enables the highest quality targets to be achieved. The customer can therefore be sure that their high-quality PCBs meet the highest quality standards, while at the same time saving costs and resources.