Machine Vision: Basics, Terms and Definitions Updated

  • October 01, 2021
  • News
Machine Vision: Basics, Terms and Definitions Updated
Machine Vision: Basics, Terms and Definitions Updated

Frankfurt, Oct. 1, 2021 - Machine Vision in industrial applications has been a success story for about 35 years and has become indispensable for measuring, inspection and positioning tasks in the production cycle in many sectors of the industry. The VDI/VDE/VDMA 2632 series of standards aims to support users and solution providers in the successful implementation of these projects. Part 1 has now been published as a draft.

An essential factor for the success of a project is a common understanding between the user and the supplier about the requirements, general conditions and the performance to be provided by the machine vision system - right from the start of the project. Misunderstandings due to unclear terms and ambiguities increase the implementation effort or even jeopardize the success of the project.

"Clear terms - comprehensive explanations: the essential basis for project success! The new version of Part 1 of the VDI/VDE/VDMA 2632 series of standards prevents expensive misunderstandings," said Dr.-Ing. Heiko Frohn, CTO, VITRONIC and member of the Board of Directors of the VDMA Machine Vision.

The aim of the revision of Part 1 was to make the technical terminology clearer and easier to understand: Definitions of terms were made more precise, and any ambiguities were resolved. For example, the terminology relating to the interaction of lighting, scene and camera system was revised and clearly visualized. The section “image evaluation” is completely new, and it has a lot to offer. Previously, the terms "reference sample," "limit sample" and "ideal sample" were used rather intuitively - a risk for misunderstandings. This is where we have provided a better clarification.

Developing a powerful classifier and configuring it in a target-oriented manner is a core competence of image processing. Recording the performance of a classifier in practice and describing it by means of key figures is another. For all those who do not work with classification and misclassification rates, false alarm rates and assignment rates on a daily basis, the new edition of VDI/VDE/VDMA 2632 Part 1 is a valuable aid for reference, as it summarizes the terms commonly used to evaluate a classification performance.

The comprehensive and precise terminology collection of VDI/VDE/VDMA 2632 Part 1 helps to avoid misunderstandings and thus unnecessary delays and costs in the implementation of image processing systems. Since the draft standard was already published in two languages (German/English), this also applies to projects with international partners.

The expert community is now being called upon to carefully review the draft. Suggestions for improvement can be submitted until Dec. 31, 2021, via objection here, so that the final standard will once again become the recognized state of the art for the next few years.

About the VDI/VDE/VDMA 2632 series of standards

The VDI/VDE/VDMA 2632 "Machine Vision" series of standards provides important assistance to suppliers and users of machine vision systems for successful joint project management. VDI/VDE/VDMA Part 2 has proven itself many times over in the preparation of requirement and system specifications. Part 3 provides important tips for the acceptance tests of machine vision systems and Part 3.1 presents methods for testing the classification performance of corresponding systems. The basis of these standards is the terminology defined in Part 1 and explained in a generally understandable way.

Did you enjoy this great article?

Check out our free e-newsletters to read more great articles..