S5-2 - Characterization of Event-Camera Biases and Operating Conditions for High-Speed Sorting
- Event
- 23. ITG/GMA-Fachtagung Sensoren und Messsysteme 2026
2026-06-09 - 2026-06-10
Nürnberg - Band
- Vorträge
- Chapter
- Event-Based Vision
- Author(s)
- P. Bäcker, G. Maier, T. Längle - Fraunhofer IOSB, Karlsruhe, J. Baßler, M. Heizmann - Karlsruhe Institute of Technology, Karlsruhe
- Pages
- 90 - 97
- DOI
- 10.5162/sensoren2026/S5-2
- ISBN
- 978-3-910600-11-9
- Price
- free
Abstract
Predictive multi-object tracking offers significant potential to improve separation accuracy in sensor-based sorting (SBS) systems. However, conventional frame-based vision pipelines are limited by exposure time, frame rate, and end-to-end latency, which reduces localization precision and actuation performance at high material throughputs. Event cameras offer an attractive alternative due to their asynchronous operation, low latency, and high temporal resolution. Yet, their performance strongly depends on sensor parameters (biases), illumination, and optical setup, and selecting suitable operating points remains a challenging multiobjective problem. In this work, we present a task-driven framework for characterizing event-camera operating points for high-speed sorting. Using a controlled rolling-sphere experiment with known geometry and an LED-based timing reference, we evaluate the influence of bias settings, illumination, and aperture on taskrelevant metrics including geometric consistency, event statistics, background activity, hot-pixel behavior, and trigger-aligned latency. We validate the stability of the proposed metrics through repetition experiments and use one-dimensional parameter sweeps to reveal distinct sensitivities and trade-offs. Based on the results, we provide guidelines for selecting suitable operating points. The presented methodology provides a basis for principled manual tuning and future automated operating-point selection in event-based sorting and related low-latency vision applications.