We’re partnering with Visometry, a pioneer in industrial augmented reality (AR), to deliver scalable, AI-based machine vision solutions for quality inspection in manufacturing. The partnership allows us to leverage Visometry’s VisionLib tracking library within our Image Twin and Product Twin offerings to provide best-in-class performance when standardizing image data and generating 3D digital twins. This will reduce model training time and scaling challenges for our manufacturer customers, helping them overcome challenges in training and scaling machine learning models for quality inspection.
Eigen Image Twin: making machine vision scalable
Eigen Image Twin is the key to unlocking the full potential of AI in machine vision at scale. This novel software tool allows manufacturers to train and deploy machine learning models faster and easier across multiple lines and factories.
By merging the images captured by the camera(s) with the CAD model of the part, Image Twin creates a digital twin that can be used to generate the same pose and perspective for all parts on all machines. This eliminates unwanted variations in image data such as different backgrounds, positions or orientations of the part. As a result, robust machine learning models can be trained with less data and scaled on all machines inspecting the same part.
“With Eigen Image Twin, manufacturers can use the same vision algorithm and AI model everywhere they produce the same part. Not only do they save time and effort in their model training and deployment, they also make their manufacturing truly smart, because new defects detected in one factory can be easily shared with all other factories around the world”, explains Josh Pickard, Director of Innovation at Eigen Innovations.
Eigen Product Twin: Turn-key vision system out of the box
Eigen Product Twin is a ready-to-use vision system for in-line assembly inspection. The kit consists of machine vision cameras with accessories such as lenses and cables and an industrial gateway with Eigen Product Twin software pre-installed.
Eigen Product Twin automatically compares the images of the part with its CAD data. If a defect such as a missing component is detected, the system alerts the staff for corrective action. It is easy to install and configure and is designed for manufacturers who want to replace manual inspection with machine vision but don’t have the time or budget for a tailored system design and integration.
“Eigen Product Twin leverages available CAD data to create a plug-and-play vision system that doesn’t require a long and complex development and installation”, summarizes Pickard.
Fast and accurate pose estimation with VisionLib
One key aspect of Image Twin and Product Twin is their ability to identify the object and its pose in the image quickly and accurately to map it with its CAD model in real-time. “During the development of Image Twin, we identified Visometry’s VisionLib as a tool to help us achieve our pose estimation goals faster”, says Pickard.
VisionLib is a multi-platform tracking library for enterprise augmented reality applications that require high-performance object recognition and precise tracking. Its Model Tracking combines CAD with image processing and brings true 3D object tracking to mixed- and augmented reality applications.
“Thanks to VisionLib and Visometry’s support, we were able to speed up our development time and achieve our accuracy and processing speed goals much faster than if we had developed the object tracking all by ourselves”, says Pickard. “We had our prototype running within a month”.
“Eigen Innovations’ approach to machine vision is groundbreaking and paves the way to the smart factory of the future. With this partnership, we are excited to contribute to delivering the benefits of AI-driven, scalable quality inspection to manufacturers around the world”, said Harald Wuest, founder and CEO of Visometry.
To speak with Josh or Harald or learn more about our software and services, drop us a line.