When your company innovation is fueled by the powerhouse of Bell Labs, there is seemingly endless capabilities at your fingertips. Many of these brilliant scientists have been doing cutting-edge research in streaming analytics, computer vision and video analytics. The capabilities are right-in-time to address the growing Industry 4.0 needs of computer vision with IoT. The way they have done it may surprise you – combining unsupervised machine learning with supervised deep learning.
The cross-roads of research maturity and market needs lined up perfectly when we decided to productize this research into Nokia Scene Analytics. Scene Analytics addresses a customer challenge of having to review massive video, audio and data streams in order to determine any unusual activity – it does this by intelligently combining algorithms with minimal pre-programming. The end result is a very specific analytics model tailored for the scene.
With cameras now being turned into IoT sensors, multiple data sources can be combined so Scene Analytics can go to work on analyzing these various streams. It learns on its own what is “out of the norm” and only records the video that is important to review. Under ordinary circumstances there would be the need for personnel to review reams of video to detect a problem, which is an unmanageable task Scene Analytics makes the unmanageable – manageable
by recording only the video that is anomalous and alerts personnel when it’s detected. Deeper forensics can also be gleaned from the IoT data, object detection and other capabilities to round out the story making the data more meaningful.
Our platform allows for customization by mixing algorithms from Bell Labs and our customers’ algorithms to meet their specific needs. For example, combining sound algorithms to also alert and record video. Check out how Room 40 is making Belgian highways safer with Scene Analytics by doing just that. This capability is a game changer that can open up so many different uses for our solution beyond safety and security.
Consider Industry 4.0 scenarios where Scene Analytics can be put to work: Think of warehouse inventory that piles up because there is process breakdown along the line. Scene Analytics could be used to detect and unusual backlog that has occurred. Customers have also asked how it can be used to protect wildlife if it were attached to a wind turbine. For example; when a flock of birds flies towards a turbine, Scene Analytics detects and alerts an operations manager to shut down the turbine. Other scenarios include optimizing operations at a port by tracking vehicle license plates and shipping container identifications. Cameras make very versatile sensors with the right algorithms – the possibilities are endless!
From protecting critical infrastructure, keeping people safe to optimizing operations, Scene Analytics is always learning. Because the algorithm is intelligent enough to know what should not be reviewed by humans, it helps reduce abuse, human bias and safeguards privacy. Allowing this software to choose the relevant events put this on a need-to-know basis.
Learn more about how Scene analytics is applied:
Share your thoughts on this topic by joining the Twitter discussion with @nokiaindustries using #IoT, #smartcities, #videoanalytics