AI-Enhanced Visual Inspection Systems for Robust Detection of Product Packaging Defects in Manufacturing Environments

Imran Ahmed*, Muftooh Ur Rehman Siddiqi, Misbah Ahmad

*Corresponding author for this work

Research output: Chapter in Book/Published conference outputConference publication

Abstract

AI-enhanced visual inspection systems are important for ensuring product quality and customer satisfaction, particularly in automated manufacturing environments where defects can disrupt supply chains and lead to contamination risks. This study applies the YOLOv7 object detection algorithm for detecting packaging defects such as scratches holes, and deformations in jar lids. The dataset contains 1859 jar lids, on average 11 per image, categorized into intact (962) versus damaged (897) jar lids. After optimizing for class imbalance and performance, the system achieved precision rates of 91.7 % overall and 3.5 % for the 'damaged' class, demonstrating robustness in challenging environments. This work presents a scalable AI-based solution to improve defect detection efficiency in manufacturing.

Original languageEnglish
Title of host publication2024 IEEE 15th Annual Ubiquitous Computing, Electronics and Mobile Communication Conference (UEMCON)
EditorsRajashree Paul, Arpita Kundu
PublisherIEEE
Number of pages6
ISBN (Electronic)9798331540906
DOIs
Publication statusPublished - 20 Nov 2024
Event15th IEEE Annual Ubiquitous Computing, Electronics and Mobile Communication Conference, UEMCON 2024 - Yorktown Heights, United States
Duration: 17 Oct 202419 Oct 2024

Publication series

NameProceedings from Ubiquitous Computing, Electronics and Mobile Communication Conference (UEMCON)
PublisherIEEE
ISSN (Print)3066-0939
ISSN (Electronic)3066-0947

Conference

Conference15th IEEE Annual Ubiquitous Computing, Electronics and Mobile Communication Conference, UEMCON 2024
Country/TerritoryUnited States
CityYorktown Heights
Period17/10/2419/10/24

Keywords

  • Deep Learning
  • Defect Detection
  • Machine Learning
  • Object Detection

Fingerprint

Dive into the research topics of 'AI-Enhanced Visual Inspection Systems for Robust Detection of Product Packaging Defects in Manufacturing Environments'. Together they form a unique fingerprint.

Cite this