Vision System Using MDP-Based Tracking for Automated Dimensional AnalysisThe NeedAccurate 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 TechnologyThis 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
Benefits/Advantages
Patent Protection
|
![]() Tech IDT2025-063 CollegeLicensing ManagerZinn, Ryan InventorsCategories |