IDS Imaging Development Systems

Inka Krischke | Inka Krischke,

AI-based image processing for the field level

Integrating AI-based machine vision tasks directly into production - no problem thanks to industrial protocols such as OPC UA and prepared function blocks of special inference cameras. And all without expert knowledge of image processing or machine learning!

© IDS

OPC UA is a key technology for exchanging data and information in industry. The industrial protocol continues to establish itself as an open standard in Industry 4.0 thanks to features such as manufacturer independence, scalability and communication security. This is also due to the fact that OPC UA works via web protocols in any TCP/IP network, making communication completely platform-independent. All of this has led to a rapid spread of the IT standard.

As a result, OPC UA has managed to penetrate and connect the entire factory. The standardized exchange across all physical layers and Ethernet-based fieldbuses up to the cloud simplifies the engineering effort enormously if data does not have to be read, converted and synchronized via different protocols. OPC UA standardizes the language between devices and defines how devices describe themselves to clients and how they are controlled. This standardizes the software-based access to camera features of any type and manufacturer and simplifies many integration processes and commissioning in vision applications today.

Advertisement

Thanks to the OPC UA shell, 'IDS NXT' inference cameras can be used directly by control system manufacturers without specialist knowledge of image processing and artificial intelligence.

© IDS

OPC UA uses so-called companion specifications, which define the most important information in order to 'describe' devices in a specific industry sufficiently and comprehensively, and thus serve as a kind of dictionary. Complex image processing tasks of inference cameras are encapsulated by OPC UA in function blocks whose exact modes of operation do not need to be known to the automation engineer or user. All necessary settings are also saved and managed in configuration data records. By replacing proprietary elements with manufacturer-independent and industry-wide accepted information structures, a simple and standardized exchange of information is created, which greatly simplifies integration processes and commissioning of new devices in the application workflow.

Self-describing inference camera

In the world of automation and industrial IoT, there are many endpoint devices that provide services and information. They act as servers in the OPC UA network. Depending on what information a client requires for further processing, it may need to connect to several servers.
Image suppliers such as industrial cameras are essentially not typical OPC UA devices. The situation is different for intelligent cameras, which have their own evaluation functionality. For embedded vision devices such as the 'NXT' inference cameras from IDS, for example, OPC UA is the perfect language for self-description in Industry 4.0. With an OPC UA shell, they can be used directly by control system manufacturers. Like a vision sensor, they work not as an image provider but as a result provider, offering information and services. They therefore act as a server in the OPC UA environment.

Product change at the touch of a button

Device and manufacturer-specific capabilities remain set via the residual interface. However, their use is effectively encapsulated by the OPC Machine Vision Companion specification and standardized as an additional interface.

© IDS

The cameras provide image processing tasks and their results via vision apps that are just as easy to use and change as apps on a smartphone. On the hardware side, the camera platform is supported by a parallel FPGA that can be programmed at runtime and acts as an AI accelerator capable of accelerating the execution of many well-known neural network architectures. The fast reconfiguration of the dedicated processor offers additional advantages in terms of future-proofing, low recurring costs and time-to-market. It is also possible to switch between several neural networks in just a few milliseconds at runtime. Thanks to OPC UA recipe and configuration management, the use of the vision apps is fully mapped in the industrial network, enabling product changes at the touch of a button, for example.

Plug-and-produce

Heiko Seitz is a technical author at IDS Imaging Development Systems in Obersulm.

© IDS

Device information and result data from the various vision apps of an 'IDS NXT' camera are disclosed to every client as an object structure that can be extended as required thanks to the OPC UA device information models and the companion specification for machine vision systems. They can navigate through this network and obtain all the necessary information - even for types that were previously unknown to them. Standardized communication via OPC UA thus creates the basis for 'plug-and-produce' new devices. This means that the 'IDS NXT' inference cameras can be used as a machine vision system in an OPC UA network with minimal effort.

Interface combination

With a REST (Representational State Transfer) web service, the cameras offer another platform-independent interface that enables simple networking with other web-enabled device classes without additional software or gateways based on the http protocol. In addition, the web service remains the decisive interface when it comes to setting up the inference tasks with the associated settings and providing them via recipes and configurations in the OPC world. This is because the OPC UA specification itself does not provide any standardized function calls for these device and manufacturer-specific processes, as image processing systems are hardly comparable in terms of their control, mode of operation and the settings and data used.

This makes it difficult to create a standard without overly restricting the functionality of the very individual systems or over-inflating a generic interface. This means that unique selling points remain possible for device manufacturers and OPC UA communication remains lean and scalable. OPC UA will therefore not replace the device interfaces, but rather supplement the possibilities in the industrial environment. Camera configurations and vision app-based inference tasks are saved in a readable format that is compatible with OPC UA.

  • Xing Icon
  • LinkedIn Icon
Advertisement
Advertisement

You might also be interested in

Advertisement

Basler

Strong annual result confirmed

Basler has presented its annual financial statements for 2020. In a declining market environment characterized by Covid-19 in the 2020 financial year, Group sales increased to € 170.5 million (previous year: € 162 million).

read more...
Advertisement
Advertisement
Advertisement

JAI

24 different models

JAI has launched the Go-X Series industrial cameras, which comprises a total of 24 different models with resolutions of 2.3 / 3.2 / 5 / 8.9 and 12.4 megapixels.

read more...
Advertisement
Advertisement
Advertisement

IDS

AI marketplace launched

On the IDS virtual AI marketplace, end customers have access to complete image processing solutions and can choose the right, ready-made application. This should open up new opportunities for image processing projects, especially for SMEs.

read more...
Subscribe to our newsletter
Advertisement
Back to home