The Ohio State University Corporate Engagement Office

Back to All Technologies

Vision System Using MDP-Based Tracking for Automated Dimensional Analysis

Engineering & Physical Sciences
Software
Industrial Processes & Manufacturing
Automation & Robotics
Algorithms
Image/Signal Processing
Platform
Standalone/Desktop Application
College
College of Engineering (COE)
Researchers
Allen, Theodore
Rodriguezbuno, Ramiro
Zhang, Yifei
Licensing Manager
Zinn, Ryan
614-292-5212
zinn.7@osu.edu

T2025-063

The Need

Accurate measurement of physical attributes, such as dimensions, alignments, and surface features, is essential for ensuring quality and consistency across industries. Traditional measurement methods, often reliant on manual tools and human oversight, are time-intensive, prone to errors, and lack the ability to provide continuous, real-time data. These limitations hinder productivity and increase costs in high-volume environments. A fully automated and precise measurement solution is necessary to address these challenges, enabling scalable and efficient quality assurance processes.

The Technology

This vision measurement system employs a single camera strategically positioned to capture comprehensive coverage of objects as they move through an inspection line. It takes approximately 30 high-resolution images per object, ensuring detailed analysis. Advanced algorithms precisely localize measurement areas, and within each area, edge detection and image processing techniques identify key points to calculate pixel distances. These pixel distances are processed through a regression model to estimate corresponding physical dimensions in millimeters. Real-time results, including dimensional control charts, are displayed on a user-friendly interface. By integrating GPU acceleration, Bayesian optimization, and MDP-based tracking (Markov Decision Processes), the system achieves high accuracy, reduces computational costs, and ensures seamless operation.

Commercial Applications

  • Manufacturing: Automated dimensional analysis for quality control on production lines.
  • Quality assurance: Inspection and monitoring of physical attributes for consistent product performance.
  • Aerospace and defense: Precision measurement of components and assemblies in regulated industries.
  • Consumer electronics: Real-time inspection of intricate features in high-tech assemblies.
  • Healthcare: Non-invasive dimensional analysis of medical devices and instruments.Benefits/Advantages

Benefits/Advantages

  • Automation and efficiency: Eliminates manual inspection and the need to take manufactured product off of the assembly line, saving time, space, and labor costs.
  • Comprehensive analysis: Captures detailed measurements through multi-image processing for accuracy.
  • Real-time data visualization: Generates control charts for actionable insights during inspections.
  • High accuracy: Combines edge detection, regression modeling, and Bayesian optimization for precise results.
  • Adaptability: Scalable to various industries and capable of analyzing diverse physical attributes.

Patent Protection

  • U.S. Provisional patent applicaiton has been filed