What’s the NXT thing in image processing?
Today, cameras are often more than just suppliers of images – they can recognise objects, generate results or trigger follow-up processes. This makes new applications possible, but on the other hand also requires different knowledge. While traditional image processing is based on a more or less inflexible code, Deep Learning turns these rule-based image processing approaches upside down. In most cases, for supervised learning that is, the desired output needs to be provided together with the data as labels to train an AI model. During the training, the AI model learns the features and rules implicitly, such that the manual effort is significantly reduced at this point. This output could be as simple as an OK/nOK statement or as complex as a semantic description of image contents.
Image processing systems are increasingly confronted with a constantly growing variety of products and variants and organic objects. Conventional approaches with rule-based image processing quickly reach their limits if the image data to be analysed varies too frequently and the differences are difficult or impossible to describe analytically. In such cases, reliable automation is not feasible due to an inflexible set of rules, even if the task is supposed to be easy for humans to solve. Through machine learning, the ability to make flexible and independent decisions can be transferred to image processing systems. Using neural networks and deep learning algorithms, we can teach a computer to see objects, recognize them, and draw conclusions from what it has learned previously. This expands the scope of application across all industries.
If set up properly, Deep Learning is a flexible and dynamic process that is capable of adapting to varying environments and temporal drifts in the scenario. However, despite these benefits, the hurdle of implementing AI-based vision applications is still quite high. This is why IDS Imaging Development Systems created the AI Vision System IDS NXT, which not only makes AI methods available, but also focuses on user-friendliness which enables its use without the need for special knowledge in Deep Learning or camera programming.
At VISION 2022, the company will present exciting developments regarding hardware and software. For example, it is not only expanding the machine learning methods to include anomaly detection, but is also presenting an additional, significantly faster hardware platform that shortens inference times by up to twenty times, can therefore execute neural networks many times faster and hence enables AIbased image analyses in the live image.
The best way to test the system and evaluate its potential for a company’s specific purposes, is the IDS NXT ocean Creative Kit. It gives users all the components they need to create, train and run a neural network. In addition to an IDS NXT industrial camera with 1.6 MP Sony sensor, lens, cable and tripod adapter, the package includes, among other things, six months’ access to IDS NXT lighthouse.
