Image credit: Relimetrics
Quality control often is a manual process, which means it tends to be inconsistent and subjective. Training people to locate and differentiate between acceptable imperfections and unacceptable flaws can be challenging, and that knowledge, once earned, often leaves the company with that employee.
Deploying artificial intelligence and Internet of Things technologies can improve quality assurance in industrial and manufacturing environments. AI-based solutions make the process more consistent, efficient, and precise.
Keep an EagleEye on the Line
Qualitas Technologies has created EagleEye, a plug-and-play deep learning vision inspection solution, powered by Intel® Xeon® processors. The Intel® IoT Market Ready Solution includes hardware, software, and a cloud-based app that integrators can deploy in an industrial environment.
Images are collected through the precision EagleEye camera, complete with multiple lighting and lens configurations, which is installed on a flexible mounting arm. The EagleEye Edge software processes the images in real time using deep learning algorithms. The EagleEye Cloud app enables image annotation, AI training, error tracing, and management. A cloud-based dashboard also enables operators to monitor accuracy and system performance. The solution requires industrial-grade edge devices, such as the HPE Edgeline servers.
The EagleEye solution has been deployed in automotive, machine, and steel manufacturing environments, but it also has been used to inspect labels, packaging, and QR codes. Qualitas claims EagleEye is ten times faster than human assessments.
Image credit: Qualitas Technologies
All Eyes on Electronics
As product customization increases, quality assurance becomes more complex, and often existing systems can’t keep up. Relimetrics, an Intel® IoT Market Ready Solution partner, is addressing that issue with a quality control solution that allows operators to train AI systems using a point-and-click interface, with no coding required.
ReliVision is an AI-based machine vision software system comprised of three components. ReliTrain allows operators to train deep learning models quickly using captured and annotated images along with “recipes” provided by Relimetrics. The system works with existing cameras and can process multiple types of image data. Once configured and trained, the AI models can be shared across plants.
ReliAudit automates visual inspections on the shop floor using the models developed in ReliVision. It integrates with existing protocols and systems, including SCADA, MES, and Modbus, using the NodeEditor. It works with industrial grade servers, and inference can take place at the edge or in the cloud. Blockchain technology ensures traceability throughout the process.
The ReliUI allows operators to review and dispute flagged anomalies. The results are used to retrain the AI models as needed.
Relimetrics partnered with Intel® to develop Reli-QA, an AI-based quality audit automation system specifically for electronics assembly. With integrated Intel® Xeon® processors, the software uses advanced computer vision and machine learning to analyze assembled products as they travel along the conveyor belt.
High resolution images are processed at the edge and compared to a predefined bill of materials. Anomalies are flagged and staff is notified, so the defective item can be pulled prior to shipping.
The system brings production ever closer to the goal of zero-defect manufacturing. According to Relimetrics, the system has a flaw detection rate of 99.9 percent and reduces audit time by 70 percent. Relimetrics also estimates that ReliVision can save a company $500,000 per line per year due to early detection of flaws and less product scrap and rework.
- Learn more at Qualitas Technologies.
- Discover more about the Qualitas EagleEye Intel® Market Ready Solution.
- Find out more at Hewlett Packard Enterprise.
- Learn more at Relimetrics.
- Discover more about the Relimetrics Intel® Market Ready Solution.
- Find out more about Intel® Xeon® Processors.