WOBURN, MA, AUGUST 10, 2026 -- BitFlow, Inc., a division of Advantech, today highlighted the use of its Axion 2xE (PCIE-1122-AE) frame grabber in a peer-reviewed badminton recognition and tracking system developed by researchers Zhihao Cui (Pingdingshan University) and Ting Zheng (Lishui University), published in the International Journal of Pattern Recognition and Artificial Intelligence.
More Headlines ArticlesBadminton presents a tough challenge for vision researchers: a shuttlecock reaches speeds up to 400 km/h (250 mph), creating motion blur blending it into the background and masking its sharp contours.
The first, a corner-point detector that locates the four corners of the court using contour and histogram analysis within a defined sensing area, achieved more than 10 times the detection accuracy of traditional corner detectors.
Coaches and players can utilize the resulting positional and flight tracking logs to break down tactical play patterns and evaluate physical fitness optimization during training loops.
That's exactly the kind of demanding, latency-intolerant application the Axion 2xE was engineered for.
WOBURN, MA, AUGUST 10, 2026 -- BitFlow, Inc., a division of Advantech, today highlighted the use of its Axion 2xE (PCIE-1122-AE) frame grabber in a peer-reviewed badminton recognition and tracking system developed by researchers Zhihao Cui (Pingdingshan University) and Ting Zheng (Lishui University), published in the International Journal of Pattern Recognition and Artificial Intelligence.
More Headlines Articles
Badminton presents a tough challenge for vision researchers: a shuttlecock reaches speeds up to 400 km/h (250 mph), creating motion blur blending it into the background and masking its sharp contours. Also, compared to a basketball or soccer ball, the shuttlecock occupies very few pixels on the camera sensor, making it prone to feature loss over distance.For these reasons, Cui and Zheng deployed two 2-megapixel high-speed cameras streaming at 250 frames-per-second (fps). To move that volume of image data off the sensors and into real-time processing without dropping a frame, the researchers built their acquisition pipeline around BitFlow's Axion 2xE, a high-performance interface board featuring a half-size x4 PCI Express Gen 2.0 bus, StreamSync DMA engine, and support for up to two independent or synchronized Camera Link cameras.The acquisition layer fed two novel detection methods the researchers developed on top of it. The first, a corner-point detector that locates the four corners of the court using contour and histogram analysis within a defined sensing area, achieved more than 10 times the detection accuracy of traditional corner detectors. The second, an elliptical detector for locating the shuttlecock's center point — built to select the correct ellipse from four candidate contours — reduced positioning error by approximately 3mm compared to conventional circular detectors.In addition, the proposed vision system enhanced tracking accuracy under complex indoor conditions, minimizing detection failures caused by swift overhead lighting changes, court background noise, and physical player occlusions. Coaches and players can utilize the resulting positional and flight tracking logs to break down tactical play patterns and evaluate physical fitness optimization during training loops."High-speed sports tracking lives or dies on the acquisition layer," said Donal Waide, Director of Business Development, iSystems, Advantech "You can build the most precise detection algorithm in the world, but if the frame grabber can't sustain a 250 fps dual-camera feed in real time, none of that downstream accuracy is possible. That's exactly the kind of demanding, latency-intolerant application the Axion 2xE was engineered for."The Axion 2xE is part of BitFlow's sixth-generation Axion family of Camera Link frame grabbers, built on a PCIe Gen 2 interface with the company's StreamSync DMA engine for high-throughput, low-latency image capture — the same class of performance increasingly demanded by sports analytics, robotics, and other real-time computer vision applications.Learn more at www.bitflow.com.
