EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

RGB‐D Vision‐Based Positioning and Anomaly Detection of Sample Plate and Vials for High‐Throughput Laboratory Automation

Kai Chen, Weidong Chen

IET Image Processing · 2026-01 · 20 (1)

DOI: 10.1049/ipr2.70335 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IET Image Processing, online erschienen am 20.03.2026. Verfasst von Kai Chen und 1 weiteren Autor:innen. Die Arbeit nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Thermo Fisher Scientific.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Chemical synthesis analytics have emerged as a cornerstone in advancing biomedical applications and bioengineering. Nevertheless, high‐throughput experimental workflows are often constrained in current automated sampling instruments, particularly in terms of real‐time measurement accuracy, geometric alignment consistency and anomaly detection. Current instruments suffer from time‐consuming vial‐to‐well coordinate confirmation, iterative height calibration and insufficient robustness in detecting physical anomalies, collectively affecting throughput and reproducibility. Driven by these challenges, this study introduces a novel measurement framework based on RGB‐D (RGB‐depth) vision fusion, enhanced by a standardized SLAS (society for laboratory automation and screening) template, enabling high‐precision, cross‐platform compatible measurement of vial position. The core innovation is the integration of 2D RGB information with 3D depth data within a geometric invariance model, allowing for rapid and accurate coordinate mapping, height measurement and anomaly detection. Specifically, the developed algorithm enables one‐shot vial‐to‐well coordinate calculation, achieving 92% precision and 93% recall, significantly accelerating the process. Furthermore, it enables automated height measurement of sample plates with an impressive accuracy of 0.52 mm, which is critical for precise alignment of the headspace sampling instrument. In anomaly detection, the 3D measurement architecture achieves 99.9% precision in identifying lean‐in vials, a 9.4% improvement over DenseSSD, the state‐of‐the‐art anomaly detection algorithm for vials in laboratory environments. Additionally, the proposed algorithm incorporates the detection of an inverted sample plate and cap presence, reducing manual intervention, minimizing contamination risks and enhancing workflow automation. This research offers a solid foundation for promoting laboratory automation. By leveraging innovative positioning and anomaly detection for sample plates and vials, it enhances efficiency, safety and reliability in high‐throughput analytical workflows.

Förderung

  • Thermo Fisher Scientific

Zitieren

APA

Chen, K., & Chen, W. (2026). RGB‐D Vision‐Based Positioning and Anomaly Detection of Sample Plate and Vials for High‐Throughput Laboratory Automation. IET Image Processing, 20(1). https://doi.org/10.1049/ipr2.70335

Vancouver

Chen K, Chen W. RGB‐D Vision‐Based Positioning and Anomaly Detection of Sample Plate and Vials for High‐Throughput Laboratory Automation. IET Image Processing. 2026;20(1). doi:10.1049/ipr2.70335

BibTeX
@article{Chen2026,
  author = {Chen, Kai and Chen, Weidong},
  title = {RGB‐D Vision‐Based Positioning and Anomaly Detection of Sample Plate and Vials for High‐Throughput Laboratory Automation},
  journal = {IET Image Processing},
  year = {2026},
  volume = {20},
  number = {1},
  doi = {10.1049/ipr2.70335}
}
Kostenlos starten →