Bin Picking With a Combination of Deep Learning and Rule-Based Algorithms

This presentation explores a hybrid approach to bin picking that integrates deep learning techniques with rule-based algorithms. Gärtner begins by introducing the concept of Deep 3D matching, which is used to accurately detect objects in a scene and compute their 3D positioning, and explains how it merges the strengths of deep learning and rule-based algorithms. The talk then covers the generation and use of synthetic training data to replace manual labeling processes that are time-consuming and subject to human error. Finally, he discusses the system's architecture, emphasizing the role of 2D area cameras and the implications for implementation in real-world scenarios.

*** This presentation premiered during the 2025 Vision Spectra Conference. For more information on Photonics Media conferences and summits, visit events.photonics.com

About the presenter

Jan Gärtner has been working at MVTec Software GmbH since 2020, initially as an application engineer. Since early 2023, he has been serving as the product manager for HALCON. He studied electrical engineering and information technology at the Technical University of Munich.