Camera innovation to help power Newtec potato processor

Qtechnology, a global developer of industrial cameras and machine vision systems, has enhanced its modular camera platform used by packaging companies including parent company Newtec.

The company’s AMD Ryzen embedded V1605B APU-based camera platform running GNU/Linux will now be powered by Hailo-8.

“Qtec is empowering industries by providing them with powerful machine vision systems used in industrial applications to solve real-world problems in a large variety of industries worldwide,” said Orr Danon, CEO and co-founder of Hailo. “With the Hailo-8, these systems will be even more powerful, ushering in a new era of smart factories powered by AI at the edge.”

Newtec Engineering A/S, which reported a 400% increase in the number of pixels processed, a decrease in power consumption and faster AI model development and deployment using libraries such as Tensorflow and Pytorch, according to a news release.

Newtec uses the Hailo AI chip in its optical sorting machines, including the Celox-P-UHD, which sorts potatoes by size, shape and quality in a single-stage process, according to the release. The machine sorts washed potatoes directly, eliminating the need for pre-grading and minimizing the need for manual labor.

With easy integration via PCIe connection, an intuitive software development kit and available Yocto recipes, the Hailo-8 ensures fast and efficient AI processing in real time while keeping energy consumption low and handling inference of high-resolution images.

With the addition of the chip, Newtec can run larger AI models in real-time while reducing power consumption.

“Hailo’s AI chip is designed to perform a large number of parallel operations,” said Aleksandar Topic, qtec software engineer. “This enables our embedded cameras to efficiently incorporate neural networks into demanding industrial processing pipelines. Dedicated components such as the Hailo-8 help us optimize how valuable resources are allocated, which ultimately allows us to perform more sophisticated analysis for our customers.”