Photonics Spectra

The use of AI to inform automated processes is a regular topic under consideration in many manufacturing settings, especially those focused on precision in high-stakes production in the automotive, aerospace, and shipbuilding industries.

AI-based optical character recognition running directly on a smart camera provides high-speed date and lot-code validation in real time, catching critical errors before products leave the factory. By Charlie Long In a typical manufacturing and distribution setting, freshly filled beverage cans are rapidly processed on a conveyor at nearly 2 m/s. Continuous inkjet (CIJ) printers fire ink droplets to mark best-by dates and lot codes onto the cans’ concave bottoms.

By Douglas Farmer Artificial intelligence continues to be implemented across the machine vision and manufacturing landscape, often carrying the promise of improved efficiency and streamlined production. However, these expectations do not always match reality. According to a report from MIT’s Project NANDA last year, many generative AI implementations had not produced measurable returns by last summer.

The choice of imaging system depends on the required level of detail, as well as the speed and cost constraints of an industrial inspection application. By Hank Hogan In industrial machine vision, the decision to use a 2D or 3D imaging system to monitor a scene depends on the type and dimensions of the information desired from the field of view. In some cases, system integrators might use both 2D and 3D to reap the benefits of a combined configuration.

Heterogeneous vision architectures based on modular edge computing platforms, informed by local AI models, provide deterministic, low-latency inference for efficient operations. By Francesco Vaiani In modern production lines, milliseconds matter for the efficient deployment of machine vision.

The editors of Vision Spectra and Photonics Spectra magazines invite you to the Inspection Summit, a two-day themed webinar series taking place July 15 to 16. This event is designed to inform machine vision engineers, imaging specialists, systems integrators, and manufacturing professionals who are developing next-generation inspection and automation systems.

Every year, more than 69 million tons of product waste are recycled in the U.S., including paper, metals, glass, plastic, and wood, according to the Environmental Protection Agency. Recycled materials are collected and delivered to a processing facility where they are sorted, cleaned of contaminants, and prepared for transport for reuse. These materials have historically been processed and sorted using rudimentary methods.

High-speed cameras and pulsed lasers can be integrated to capture transient changes during welding and laser processing, providing insights to improve methods. By Ethan Fu Amid the rapid advancement of high-end manufacturing, welding — a critical joining process — directly affects the reliability and market competitiveness of end products, such as power batteries and automotive engine blocks, through its quality and efficiency.

STOCKHOLM, June 25, 2026 — Hexagon AB, a digital reality solutions company, acquired ITRES Research Limited, a Canada-based provider of imaging systems. The acquisition strengthens Hexagon’s portfolio with ITRES’ high-performance hyperspectral and thermal imaging sensors, giving the company a more complete multi-sensor mapping platform.

indie Semiconductor, a developer of automotive sensing solutions, has signed a definitive agreement to acquire the fabless CMOS image sensor group from ams OSRAM for €40 million ($47M). The transaction — expected to close in the third quarter — is the latest divestment from ams OSRAM; The company said in an announcement earlier this year that it planned to examine potential sales of its business units. The company subsequently announced the sale of its non-optical sensor business to Infineon.