How video and motion analysis can help prevent machine downtime
Methods to visualize events and track physical movement are becoming more sophisticated, helping to reliably detect early signs of machine failure and identify suspicious activity or intruders. Today’s motion and video analysis technologies are smarter, faster, and provide more context for investigation, decision making, and corrective actions.
Motion analysis
Vibration analysis cameras provide a non-contact means to monitor everything in view for vibration problems. Iris Edge from RDI Technologies brings Motion Amplification with its continuous full-field, camera-based vibration measurement into the IoT world. Using the MQTT open standard, it allows integration of the data into third-party processing systems such as SCADA, HMI, ML/AI, and the cloud.
“Iris Edge, with its pan/tilt and Power over Ethernet (PoE) capability, enables full-field measurements over enormous fields of views of the running equipment, but also unlocks more assets such as structure, supports, piping, foundations – items that typically go undetected yet may be at the root of a critical problem," explains Jeff Hay, CEO of RDI Technologies.
Metavision neuromorphic sensors from Prophesee support vibration monitoring and frequency analysis for predictive maintenance as well as other monitoring, safety, and inspection applications. A single HD Metavision sensor can provide a million independent measuring points and a “holistic vibration monitoring understanding, meaning many parts or machines can be viewed,” says Artiom Tchouprina, senior application engineer at Prophesee.
Specifically, a Metavision sensor “can measure vibration frequencies from Hz to kHz – simultaneously in the same scene – continuously, remotely, with pixel precision, by tracking the temporal evolution of every pixel in a scene. For each event, the pixel coordinates, the polarity of the change, and the exact timestamp are recorded, thus providing a global, continuous understanding of vibration patterns,” Tchouprina adds.
The new AI-based MultiSensor intrusion detection system from Senstar combines video analytics and motion sensors to improve detection and reduce nuisance alarms. Its embedded sensor fusion engine synthesizes data from multiple sensing technologies, from an image sensor with 180-degree field of view to short-range radar, accelerometer, passive infrared (PIR), and high-frequency vibration sensors.
“Conventional sensor technologies focus on detecting specific physical phenomena, and by themselves do not take into account the full context in which an event is occurring. With Senstar MultiSensor, we can provide complete and precise situational awareness for each event to make the best decision,” says CEO Fabien Haubert.
Video analysis
Video management software from Qognify (part of Hexagon) now supports greater video flexibility and availability and a wider range of analytics. The latest release, VMS 7.4, includes a new Qognify Web Client and improved forensic search tool, and makes available for purchase a suite of smart video analytics powered by SAIMOS technology including perimeter security, counting, object detection, fire/smoke detection, face recognition, and gender/age analysis.
VP of Products Sunil Mudholkar says, “Qognify VMS 7.4 empowers organizations to detect, understand, and respond to security incidents quicker and more efficiently. A new web client gives users the ability to access relevant information when they need it, wherever they are. And new AI-based video analytics capabilities generate and leverage metadata to deliver powerful insights for aiding investigations.”
Deep learning video analytics from Axis Communications optimize camera image quality and AI performance. For instance, its Forensic Wide Dynamic Range (WDR) reduces visible noise and artifacts. Lightfinder technology delivers high-resolution, full color video with minimal to no motion blur, even in near darkness. OptimizedIR combines camera intelligence and LED technology so that the entire field of view is always evenly illuminated.
“We ensure our deep learning systems are compatible with the wide range of products Axis offers, from fixed cameras to fisheyes, intercoms, and other solutions. As a result, customers can choose a form factor that works for their specific situation without sacrificing analytics performance,” explains Steve Burdet, manager of solutions management at Axis Communications. “Additionally, we are continuing to improve ML performance on older systems to close the gap with modern deep learning devices.”
For video-based smoke and flame detection, the new Aviotec 8000i IR camera from Bosch Building Technologies leverages AI algorithms based on deep learning and an integrated infrared illuminator to distinguish real fires from false alarms at the source within seconds. It can operate 24/7, even in complete darkness, low visibility, and environments subject to wind, dust, or humidity.
The Aviotec 8000i IR reportedly outperforms aspirating smoke detectors with 3x faster detection times on average. It is a stand-alone unit including an IP67 waterproof housing, motorized lens, infrared illuminator, and 4-megapixel sensors.
