Intelligent Monitoring System for Carton Packaging Defects Based on Computer Vision
Abstract
Intelligent Monitoring System for Carton Packaging Defects Based on Computer Vision
Incoming article date: 14.01.2026Introduction. Ensuring the quality of cardboard packaging is a critical challenge for modern warehouse logistics, as damaged packaging increases the risk of product loss and negatively affects customer satisfaction. With the rapid growth of e-commerce, there is a growing need for automated and reliable quality control solutions based on computer vision technologies.
Aim and objectives. The aim of this study is to develop an automated monitoring system capable of detecting and classifying defects in cardboard boxes in real time under warehouse conditions. The objectives include designing a defect detection model, integrating it into a web-based system, and evaluating its performance in practical scenarios.
Methods. The proposed solution is implemented as a web application that integrates a YOLOv8-based deep learning model trained on a custom dataset of defective and intact packaging images. The backend is built with Flask for real-time video stream processing, while Apache Superset is used to provide analytical dashboards for visualizing defect statistics.
Results. Experimental testing in storage, sorting, and delivery scenarios demonstrated high detection accuracy exceeding 95% and stable performance under varying lighting conditions and partial occlusions. The system successfully identified major defect types such as dents, tears, and deformations with minimal false positives.
Conclusion. The developed monitoring system proves to be an effective tool for improving packaging quality control in warehouse operations, reducing operational risks, and supporting data-driven management decisions in logistics environments.Keywords: computer vision, defect detection, carton packaging, YOLOv8, deep learning, monitoring system, video analytics