The real benefits of automated quality control in production lines can only be realized when companies implement 'intelligent' automation with functions such as smart data. Thanks to image processing algorithms such as edge-based sparse features' variation absorbing templates and fast, parallel hardware architecture, today's systems make it possible to achieve a detection speed with optical inspection systems that is over ten times faster than conventional inspection systems without new algorithms. Compared to older algorithms, some of which were error-prone and inaccurate, these systems work over 100 times faster and increase detection quality at the same time.

In the 2000s, edge-based algorithms led to improvements in light changes and shadows, but were still inaccurate in terms of sharpness and contrast. The new 'Sparse Edge Detection' algorithm takes the information used and reduces it to clearly identifiable and representative points. This minimizes errors and at the same time significantly improves speed.

Recognize complex patterns

Artificial intelligence (AI) is an important trend in this context and with regard to advanced vision systems in combination with deep learning algorithms. For the food and beverage industry, for example, this can mean that the inspection software is trained using an algorithm to recognize complex patterns for the respective product. In this way, for example, irregularly shaped objects such as baked goods can be easily inspected and even defects that have not previously been explicitly taught can be detected. Such AI-supported inspections can relate to the shape, color and texture of a product.

Conventional vision systems can be programmed to find specific defects and detect deviations, for example. However, an AI-based inspection system can learn to detect abnormal products without being specifically programmed for every deviation from the norm.

Another approach is to recognize and read OCR markings. In very demanding applications, the letters may not match the pattern of a traditional software setting without AI. When reading OCR with image processing software, the traditional software without AI may not be able to interpret which letter it sees. With AI, a more robust and reliable reading can be achieved.

Another aspect that is important in terms of quality control is the connection to a database. It enables the recording of quality inspection and production data as well as ensuring traceability and compliance with legal regulations. Data analysis can be used to identify trends and optimize predictive maintenance. An AI-supported machine controller can offer this connectivity while combining logic, motion, vision, safety and robotics in one solution. For example, a controller can take over communication with several smart cameras and register the results directly in the SQL database via an internal SQL client. It is important to ensure that the controller can communicate directly with SQL databases by using integrated function blocks that provide SQL queries. Industrial PCs can be used to visualize the results at any station in the factory.

The system solution

Many production lines have special quality control requirements. In the pharmaceutical industry in particular, errors must be avoided at all costs. For this reason, production lines in this sector are more comprehensively protected against defects than in any other. And even if the consequences may not be as serious in other industries, errors and defects are more or less not allowed to occur there either. For companies that want to position themselves in a future-oriented and compliant manner and also want to approach the topic of quality control strategically and comprehensively, it is advisable to choose a holistic system that covers all production line tasks, including quality control.

This is possible with systems from Omron, for example. With these systems, the data transmitted by a vision system is processed on site and made available centrally via the cloud for detailed analysis so that appropriate measures can be taken. The complete networking of the systems ensures a better connection between the machines on a production line.