Author: James Butcher, IOTech Product Manager
One of the hottest subjects in IoT and edge computing right now is how advanced AI techniques such as video inference and computer vision technology are moving towards the edge. Obtaining immediate local insights from vision-based computing can revolutionize how businesses run their operations. There are many use cases such as predictive maintenance, product quality checking and theft detection that become far more accurate when video inference is involved.
Nothing Without the Cameras!
While video inference technology is becoming lighter and more suitable for edge computing, don’t forget that in order to perform any computer vision operations you must first be able to securely access a video stream from the onsite cameras.
Standardization is Key
There are hundreds, if not thousands, of different camera models and manufacturers on the market today. Cameras range in capability from simple static surveillance cameras through to advanced equipment that can stream, record and even run onboard video processing operations. It can be an integration headache to connect and manage these cameras but fortunately there are international standards that help address this.
One of the main camera standards is ONVIF which stands for the Open Network Video Interface Forum. Many companies are choosing ONVIF-based cameras because you can connect and stream data from them in a standard and consistent way. You are not locked to specific cameras suppliers while site installations and maintenance becomes much more predictable.
ONVIF Camera Connector
As specialists at the IoT Edge, IOTech’s software products allow users to connect, onboard and acquire data from a whole host of different OT devices. To add cameras to that list, IOTech have created a software service that lets you easily connect to and control ONVIF compatible devices. Our ONVIF connector was released as part of Edge Xpert 1.8 and was recently updated to include support for remote control of the cameras via PTZ (pan, tilt & zoom) operations.
So with this connector, an operator can fully control the cameras in an automated way and gain access to the video feeds required for the computer vision engines mentioned above. The result of that analytics, e.g., the number of people in a room, or a positive product identification, can then be used in decision making and control operations at the edge via Edge Xpert or Edge XRT. Stay tuned for details of upcoming IOTech software that can also run integration to the vision inference engines in a standard way.
One of our engineers, Amaan Akram, has taken our ONVIF connector and produced a cool demo video that shows how the service can be used to onboard, control and access the video feeds from a couple of ONVIF cameras running in our lab:
More Info and Download Now
Available as a connector for Edge Xpert, EdgeX Foundry and shortly also Edge XRT, the ONVIF connector is ready to be used for your edge camera control needs.