Reliable information system for identifying spatio-temporal continuity of kinetic deformed objects with big point cloud data

Claire Y. T. Chen, Edward W. Sun () and Yi-Bing Lin
Additional contact information
Claire Y. T. Chen: Montpellier Business School
Edward W. Sun: KEDGE Business School
Yi-Bing Lin: National Yang Ming Chiao Tung University

Annals of Operations Research, 2025, vol. 349, issue 1, No 7, 103-138

Abstract: Abstract In the context of Industry 4.0, a wide range of sensors are extensively deployed to gather production and equipment operation data, while also connecting human workforce information through the industrial Internet of Things technology. This integration enables effective improvements in sustainable, human-centric, and resilient productivity by leveraging industrial process control and automation. In this paper, we propose an intelligent information system for analyzing large point cloud data sets from depth sensors, which are used for detecting, representing, locating, and shaping monitored objects. To address privacy concerns, our system only considers de-identified information during analysis, using a newly proposed dynamic clustering method based on multivariate mixture Student’s t-distribution for monitoring human motions. The information system consists of two main blocks: segmentation and dynamic clustering for monitoring or tracking. The segmentation algorithm, utilizing a multivariate mixture Student’s t-distribution, groups points into homogeneous partitions based on spatial proximity and surface normal similarity, without relying on any semantic indicator or pre-determined shape. The dynamic clustering algorithm, powered by an online learning state-space model, efficiently incorporates and updates the centroid position and velocity of the object being monitored. To evaluate the reliability of our proposed method, we introduce two time-consistent measures that account for different illumination levels, drastic limb movements, and partial or full occlusions during object motion processing. We conduct empirical experiments using a large point cloud data set, comparing our method with several alternative methods. The results highlight the superiority of our proposed method.

Keywords: Machine learning; Big data; Dynamic clustering; Internet of things (IoT); Depth sensor; Information systems (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://link.springer.com/10.1007/s10479-023-05522-z Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:annopr:v:349:y:2025:i:1:d:10.1007_s10479-023-05522-z

Ordering information: This journal article can be ordered from
http://www.springer.com/journal/10479

DOI: 10.1007/s10479-023-05522-z

Access Statistics for this article

Annals of Operations Research is currently edited by Endre Boros

More articles in Annals of Operations Research from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().