A Vision-Based Smart Laboratory Asset Management System with Edge-Cloud Architecture

Laboratory operations in academic environments are frequently managed through manual logbooks, ad-hoc spreadsheets, or isolated digital tools, leading to incomplete asset traceability, inefficient scheduling, and increased administrative burden. This paper presents the design and implementation of a vision-based smart laboratory asset management system that integrates edge-based perception with a cloud-hosted backend to automate equipment identification, borrowing and return logging, and laboratory booking workflows. The system employs a deep-learning object detector for real-time recognition of laboratory components and uses QR code decoding as a deterministic validation mechanism to reduce misidentification under operational conditions. A layered edge-cloud architecture is adopted in which an edge node performs image acquisition and inference, while centralized backend services enforce authentication, role-based access control, inventory consistency, and audit-ready transaction logging. The paper details system architecture, module decomposition, control flow, data modeling, deployment considerations, security governance, and system-level evaluation, providing a practical blueprint for deploying vision-enabled laboratory management systems in academic environments. System-level evaluation shows stable operation under varying illumination and pose conditions, with QR-based validation introducing an average overhead below 0.25 s.