n TECHNOLOGY September 2026 www.drivesncontrols.com 16 SICK HAS ADDED AI deep-learning algorithms to its Nova 3D machine vision platform to allow quick customisation, and open up new capabilities, including height analysis. Sick says that the AIpowered 3D image processing goes far beyond the human eye to ensure faulty products are removed from production and packaging processes at an early stage. The AI capability allows configurable 3D machine vision, and improves quality control using intelligent inspection. An example-based approach, combined with on-device training, simplifies the development of vision applications and offers a cost-effective way of incorporating deep learning into quality control, without needing any extra equipment. A 3D height analysis capability unlocks advanced quality-control applications that Sick says were previously difficult or impossible to achieve with traditional rules-based inspection. Key applications include: n inspecting deformed packages in sectors such as food & beverage and consumer goods; n detecting empty boxes or missing objects in totes or matrix packaging, in fulfilment and logistics operations; n verifying assembly and inspecting surfaces in electronics and battery production; n inspecting and classifying in automotive applications; and n checking completeness in manufacturing operations. For example, AI-powered inspection of deformed packages can validate the appearance, integrity and function of packaging after filling, measuring deviations in the shape and surfaces of packages. It can detect defects at high speeds, even if there is no colour contrast. Teach-by-example using sample images simplifies set-ups. The AI is trained using an organisation’s own data. Users collect data, train models and execute inspection tasks directly on the device, with quick set-up of inspection tasks, allowing rapid batch changes. Every pixel carries a height value, enabling 3D data and defects to be reconstructed by a trained neural network in the Nova AI software. It allows colour- and contrast-independent inspection and improves quality control, with defects being displayed on an anomaly heatmap. The toolset also includes traditional rules-based machine vision software tools. “Our latest innovation combines Sick quality, expertise and consistency with deep learning to meet demand in two of the highest growth product categories in machine vision: AI software and 3D imaging,” says Sick product manager, Diego Quintana Tukasaki. “Introducing neural network AI capabilities into our high-performance Nova platform expands 3D inspection beyond its traditional limits. “Seeing far beyond the human eye, it provides greater control over data and analytics to deliver highly accurate, configurable, easy-to-train, customised intelligent inspection solutions,” he adds. “Inspected items fulfil required quality and sorting demands, helping to improve yield, reduce waste and increase customer satisfaction, while problems with batches or production lines are swiftly identified with the anomaly heatmap.” www.sick.com/gb/en TELEDYNE DALSA HAS ANNOUNCED a “groundbreaking” shortwave infrared (SWIR) hyperspectral camera that, it says, redefines the cost-to-performance ratio for this type of inspection. The Kaleido camera has been designed to overcome the integration and performance barriers of industrial sorting, improving material discrimination and increasing throughput “significantly”. The camera delivers a spatial resolution of up to 1,280 pixel and line rates exceeding 2.3kHz, allowing more items to be inspected in less time, without sacrificing spectral precision. The camera, which incorporates a sensor, spectrograph and interface, is aimed at critical applications in recycling, food safety, pharmaceuticals and waste management. Built‑in spectral band selection and spectrograph‑based aberration correction simplify its use, ensuring consistent calibration across the operating temperature range. These innovations are claimed to accelerate time-to-market for machine-builders. Teledyne has integrated proven functions from its linescan products, such as metadata tagging and multiple regions of interest, ensuring that only the most relevant data is delivered to the processing engine. The camera has a 10 GigE interface that complies with the GigE Vision and GenICam standards. Its sensor maintains strong signal integrity even in low‑light conditions, allowing energyefficient SWIR LED lighting to be used. The camera “combines deep sensor physics, advanced optical design, and world-class system engineering to reduce the complexity and cost of hyperspectral deployment,”says Sadiq Panjwani, senior executive vice-president and general manager of Teledyne Dalsa’s machine vision cameras group. “Kaleido is a scalable solution that improves global outcomes by ensuring food safety, driving cleaner manufacturing, and enabling more responsible resource management. This is the future of industrial intelligence.” www.teledynevisionsolutions.com High‑speed hyperspectral inspection camera overcomes sorting barriers AI-powered 3D machine vision system combines deep learning with height analysis Sick says that its AI-powered 3D machine vision system goes far beyond the capabilities of the human eye
RkJQdWJsaXNoZXIy MjQ0NzM=