Feature: Machine vision
In this new era of unsupervised
In this new era of unsupervised or semi-supervised learning, the initial training of the model can take place on the production line, completing in just 24 hours before being ready to run
The first stage Deep learning has now become the leading way of providing quality inspection. These solutions solve complex problems that were hard to handle with rule-based systems, like detecting scratches or dents on surfaces that are poorly defined. By receiving many examples of defects, a deep-learning model can generalise the problem and detect defects on new parts by deploying this more general understanding. However, a problem ensued: the
model required training images to be entered manually, a very lengthy and difficult process. During model training, many images must be reviewed first. For example, if the model fails to accurately detect a crack, more samples are required, until the model understands what a crack is. Furthermore, simply providing the images is not enough, as some mark-up around the images is also needed. For example, you might need to draw a box around the defect, or some other way of segmentation, tracing the exact route of the crack. It is important to bear in mind that
the model can only be trained with images of defects. The end user must review thousands of images, identify the ones with defects, then carry out the correct mark-up before providing it to the machine. For most deep-learning applications,
hundreds if not thousands of examples are needed, just to get started. If you are manufacturing 100,000 parts a day, and you know that 2% of the parts you produce are defective, that’s many images to review.
One approach to avoid this headache
is to feed the model man-made images of defects. In other words, the customer will manually create defects and feed those images to the model. Some companies have even sought
to develop software that generates images of potential defects, but both approaches run into the same problem: Artificially-created defects simply can’t accurately represent defects that naturally occur in the production process.
Unsupervised learning: a new era? We are now standing at the beginning of a new, third era in quality inspection. Technological solutions aimed at overcoming the limitations of previous deep-learning solutions are generally referred to as “unsupervised” or “semi-supervised” systems. A key goal is to automate the process of model building described earlier. Whereas traditional deep-learning
solutions require examples of defects for training, a semi-supervised model can be fed images of a sample of non- defective parts, so-called “OK parts”. The initial model, whilst not perfect, will provide a basic understanding of what constitutes “okay”. It will then use this generalised understanding to flag suspected defects or outliers and, with feedback from the customer, it will continually update and learn. Automating the model-building
process in this way is much easier for the user, but the biggest return on investment comes from the reduction in the time it takes to have an effective model that works.
learning, the process of model building can take place on the production line. Previously the process of data categorisation would require one person working on their own for at least a week before returning to the production line with a model. After applying the model, they would often discover it doesn’t work as intended, requiring further training, which adds interruptions to the production. In this new era of unsupervised or semi-supervised learning, the initial training mode can take place on the production line and can be complete in just 24 hours before it is ready to run. A semi-supervised solution is arguably
superior to a fully-supervised one, since the outcome is a system that better understands the product. Until now, many of the people working on training models were experts in data and AI but lacked a good understanding of the product. In this new era, semi- supervised systems will leverage the knowledge of the production people, with their feedback optimising the model. A final advantage for those willing
to embrace this new era is the ability to differentiate between defects. Previously the only goal was detecting a defect and classification was less important. However, more advanced deep-learning solutions can cluster together different images and build an understanding of different types of defects. The data gathered from this can then help support both preventive and predictive maintenance.
AI-enabled In any field of technological progress, we enter new eras where the changes that occur are not just at the margins, but amount to a trend shift. Thanks to AI developments, we are entering a new era where problems previously thought too complex for deep-learning solutions are now becoming solvable, and the difficult task of building a workable model is a lot quicker and easier thanks to automation.
www.electronicsworld.co.uk July/August 2023 45
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60
