Point cloud trueness - the secret to reliable picking

This is a 30-minute webinar with approx. 10 mins available for Q&A afterwards.

• Part 1 with Øyvind Theie, VP Products - Zivid: Trueness, the forgotten metric.
• Part 2 with Nicolò Boscolo, Product Manager - IT+ Robotics: Detecting and Picking with confidence.

3D machine vision systems are used extensively across a multitude of industries including manufacturing and logistics. Consistently reliable identification and picking of objects are absolutely necessary. But it doesn't always work out that way, and missed picks and even robot stalls can occur. A common cause is a lack of 'trueness' in the 3D vision system.

Accuracy, precision, and trueness are critical factors in 3D machine vision together form the quality of the camera's perception of reality. Zivid 3D color cameras are designed to deliver astonishing levels of trueness. But there is still confusion around these topics.

So we will take a deeper dive and explore what these things are, how they occur, and how we measure them. We will look into the best techniques used to get 'true-to-reality' images for robots and machines to work with and show why Zivid 3D color cameras excel at 'trueness'.

Topics we will cover:

• The cost of unreliable picks
• Why do picks fail?
• Aligning what the camera sees with reality
• Accuracy, precision, trueness, and their relationship.
• Field-of-view and the 'sampling grid'.
• Precision and 'seeing the details'.
• Hand-eye calibration techniques
• Trueness - 'True-to-reality'.
• Zivid 3D Color camera systems.
• Conclusion and summary
• Q&A.