ISBN (online): 978-83-8419-053-1 OAI
DOI: 10.18276/978-83-8419-053-1-31
.
TOWARD INTELLIGENT PALLET POOL MANAGEMENT – PALLET DAMAGE IDENTIFICATION AND INPUT DATA STRUCTURE IN CAMERA- AND AI-BASED DECISION SUPPORT SYSTEMS
| Authors: |
Mariusz
Sowa
Department of Logistics, Institute of Management, Faculty of Economics, Finance and Management, University of Szczecin, Poland Katarzyna Kędzierska Department of Technological Processes, Faculty of Economics and Transport Engineering, Maritime University of Szczecin, Poland Joanna Tuleja Department of Technological Processes, Faculty of Economics and Transport Engineering, Maritime University of Szczecin, Poland |
| Keywords: | EPAL wooden pallets Structural damage Logistics operations Intelligent vision systems Industrial AI Machine learning Damage classification Automated inspection decision-making |
| Whole issue publication date: | 2025-10-02 |
| Page range: | 13 (469-481) |
| Klasyfikacja JEL: | C44 C63 D22 D24 L15 M11 O33 |
Abstract
Purpose: The aim of this article is to conduct a comprehensive assessment of structural damage in reusable EPAL wooden pallets (800 × 1200 mm) and explore the potential of intelligent vision systems as tools supporting technical diagnostics and decision-making in pallet pool management.
Need for the study: Given the essential role of pallets in ensuring continuity and efficiency of logistics chains, there is a growing need for reliable and objective methods to assess their technical condition. Current practices based largely on visual inspections are time-consuming, error-prone, and lack standardization.
Methodology: The study involved observational analysis of 135 pallets made from both dry and wet structural components. The inspection environment was carefully prepared to ensure uniform lighting and ergonomic station layout, enabling precise identification of surface defects and safe handling. All types of damage were photographed and categorized, forming a consistent classification system.
Findings: Significant differences were observed in the types and frequencies of defects depending on material moisture. Dry-component pallets exhibited primarily cracks (47.19%), construction gaps (16.85%), and warping (13.48%), while wet-component pallets were mostly affected by discoloration (39.24%), nail corrosion (29.11%), and resin leakage (15.56%). These findings highlight the impact of production parameters on pallet durability prior to operational use.
Practical Implications: The developed damage classification system can serve both traditional technical diagnostics and training datasets for machine learning models. Integrating intelligent vision systems with Warehouse Management Systems (WMS) enables automated, fast, and consistent defect detection, maintenance planning, and withdrawal of damaged units. Neural network-based solutions reduce analysis time and eliminate human subjectivity. This approach supports the digital transformation of logistics in line with Industry 4.0 principles and complements expert-based methods with AI-powered efficiency.
Download file