Manufacturers must contend with high turnover, as well as yield, traceability, manual inspection costs, and shrinking margin concerns. Conventional visual inspection is not only time-consuming and costly, but also prone to human error. Advantech Co. and Overview last month joined forces to offer an “end-to-end” artificial intelligence system for vision inspection.
Founded by Tesla automation engineers, Overview develops platforms using deep learning to automate and improve inspection processes. The San Francisco-based company said its cameras and next-generation AI can improve the speed and reliability of inspections while saving companies significant costs. Overview said it has designed its product specifically for flexibility and ease of use, allowing accessibility to all manufacturers, regardless of size or specific use cases.
“We started Overview in 2018 because after years of building automation equipment for Tesla, we became frustrated with the experience of trying to implement vision systems for quality checks,” wrote the company in a blog post. “Vision to us meant $30,000 in hardware, no connectivity, an ancient user interface, and a requirement of steep technical knowledge and training.”
“It also meant that you could write a 200-line algorithm to handle the variance in your process, only to have it totally break under the slightest change in the product or environment,” said Overview. “It also meant being capped at mediocre accuracies with no obvious path forward—other than a complete hardware redesign or fixing a spiderweb of algorithm code.”
“We believed that a combination of deep learning—which is superior at handling variance and achieving high accuracies—with the ability to access a vision expert with a single mouse click would fix this,” the company added. “We knew, however, that software alone cannot solve difficult problems in manufacturing.”
Edge camera plus deep learning equals quality
Founded in 1983, Advantech is a leading provider of industrial Internet of Things (IIoT) and edge hardware technology. The Taipei, Taiwan-based company said its IIoT hardware and software can assist business partners and clients in connecting industrial chains. Advantech, which has U.S. offices in Irvine, Calif., also works with domain-focused partners to co-create IIoT systems.
The combination of Advantech's new ICAM-500 edge AI camera and Overview's deep learning technologies will help manufacturers improve quality control, traceability, and speed for production-line operations, said the companies.
A reliable, automated visual system can review product assembly at each stage of the manufacturing process, which simplifies and improves product inspection, the partners said. AI and automation can eliminate the complexity of finding good and bad products, leading to zero defects in the manufacturing process, claimed Advantech and Overview.
“Together, we deliver more imaging and AI horsepower than existing vision systems with a far better user experience, and for a fraction of the price,” said Overview.
Overview Snap Platform handles multiple tasks
Overview said its Snap Platform from Overview is an adaptable machine vision system that handles everything from device management and algorithm development to product quality support and traceability. When integrated with smart cameras such as the Advantech ICAM-500, the Snap Platform can automatically inspect from a programmable logic controller (PLC) trigger, interval, or—in the case of video events—continuous capture.
“We are continually evolving our Snap Platform to provide new and existing customers with easy-to-use, powerful inspection systems that also grow and scale with changing business and quality needs,” said Chris Van Dyke, CEO of Overview. “By partnering with best-in-class hardware providers for a reliable, end-to-end solution, we make device setup easier and faster. We are excited to bring Advantech into the fold and integrate our Snap platform with its high-quality, industrial edge AI devices.”
Unlike a typical vision system, the Snap Platform saves all relevant process data and creates a traceable visual record of every unit, said Overview. It catalogs data for easy search and analysis.
This helps with traceability, as well as the discovery of yield loss sources, the company said. Users can determine the root cause of losses with data post-processing.
Advantech uses NVIDIA modules in camera
Embedded with NVIDIA Jetson Nano modules, Advantech’s new ICAM-500 AI camera combines an industrial-grade image sensor, advanced LED lighting, and a variable focus lens with acquisition and AI computing capabilities.
The company said these features enable systems integrators and software vendors to integrate AI and machine vision into applications. They can also efficiently perform AI automated optical inspection (AOI), AI optical character recognition (OCR), and object recognition at the edge.
“The ICAM-500 is a unique device, as the combination of NVIDIA computing modules and camera system offers image acquisition and AI inference functionality all within the same system,” said Carolyn Swan, director of IoT partnerships in the Advantech IIoT Group. “Integrating the ICAM-500 with the Overview Snap platform creates a state-of-the-art solution for the manufacturing industry.”
“Our built-in camera helps reduce latency challenges that usually occur due to the distance between IP cameras, the cloud, and AI inference systems,” she said. “That low latency improves efficiency of on-site AI inference, creating an ideal solution for edge AI applications on the production line.”
Advantech and Overview said customers can deploy and manage their combined system without high-level technical expertise. Overview experts support the Snap Platform remotely, so customers do not need on-site machine learning experience.
It also integrates with existing factory systems and communicates across networks to accomplish more efficient, lean, and profitable operations, said the company. The Advantech ICAM-500 is available now.
Article topics
Email Sign Up
Related Inspection News
Related Companies
Related Resources
