Computer VisionActive

Inventory - Computer Vision-Based Shipyard Inventory Management

Novel computer vision and probabilistic inference system for tracking 600,000+ sheet metal parts in shipyard with IoT integration

inventory-managementcomputer-visionshipyardiotmanufacturingtracking-system
Inventory - Computer Vision-Based Shipyard Inventory Management

Project Overview

Our computer vision-based shipyard inventory management system tackles the monumental challenge of tracking over 600,000 sheet metal parts across vast industrial facilities where traditional RFID and barcode systems fail due to harsh environments, metal interference, and scale limitations. The project integrates multi-camera computer vision networks with probabilistic inference algorithms and IoT sensor fusion to automatically identify, locate, and track parts as they move through cutting, storage, and assembly operations. The core technical innovation combines robust object detection models trained on industrial environments with Bayesian tracking algorithms that maintain confidence-scored location histories despite occlusions, lighting variations, and cluttered backgrounds. Key engineering challenges include developing scalable vision systems that process thousands of simultaneous objects, creating probabilistic models that handle uncertainty in harsh industrial conditions, implementing real-time distributed computing across multiple camera nodes, and ensuring system reliability in environments with extreme temperatures, vibrations, and electromagnetic interference.

This revolutionary technology transforms manufacturing operations by providing unprecedented real-time visibility into massive inventory systems, enabling instant part location, automated audit capabilities, and data-driven production planning that eliminates costly search delays and inventory losses. The system dramatically reduces manual labor requirements while improving accuracy compared to traditional tracking methods, supporting just-in-time manufacturing, optimizing workflow efficiency, and preventing the multi-million dollar losses common in large-scale industrial operations. Applications extend beyond shipyards to aerospace manufacturing, automotive assembly, heavy industry, and any large-scale manufacturing environment where part tracking complexity overwhelms conventional systems. Our research team seeks partnerships with major manufacturing companies for deployment validation, technology companies for commercial scaling, and IoT providers for sensor integration, offering proven expertise in industrial computer vision, distributed systems, and manufacturing process optimization.

Project Videos

Video

Team Members

PKS
Professor Kenji Shimada

Project Details

Started

January 10, 2023

Category

Computer Vision

Status

Active

Interested in Joining?

We're always looking for passionate researchers to join our team.

Contact Us