Computer Vision – Grand Dad of All IoT Sensors

Human body is the most efficient and powerful system that exists and is a result of evolution over millions of years. And it would make complete sense to emulate its mechanism as we build modern systems.

IoT or “enchanted objects” as I like to call them are also moving the same direction. The 5 human senses that operate through the organs map to the nodes in an IoT system. Same is the case of actuators that map to the voice, gestures, movement and other actions that humans perform through respective actuation organs.

Spinal cord maps to the IoT gateway that forms the backbone of any IoT system and does not just serve the purpose of communication between the organs and the brain but also provides the reflex actions through a local intelligence that has been driven back from the brain by the models created through experiences and DNA

And of course the brain that maps clearly to the cloud platform in an IoT architecture. We can go on an on with this analogy, but that is not the focus of this blog.

Now that we have established that its the human body that we are trying to emulate, lets understand which is the most powerful sense of the human body. Vision. Right? If that is the case, wont it make complete sense[pun intended] to use the same sense effectively while building an IoT system? Lets park this thought for the time being while we evaluate the challenges with current sensors.

When it comes to IoT solutions, getting the right sensor with accurate reliability is the toughest part. That we have learnt through numerous challenges that we have faced with even simplest of use cases. Retrofitting the sensors or reading information from existing sensors [analog or digital] is the other challenge because bulk of use cases we come across require IoT enabling existing assets and systems.

Significant investments are required to update infrastructure, get the sensors installed, create complex networks and its a multi year job to get it all correct and operational. Then comes the challenge with the wear and tear of these sensors and they need to be continuously calibrated or maintained.

Vision has huge power and images / video is our answer to it. Only if we can tap into the information that it provides, we can create some really powerful IoT systems. We can “read” existing analog dials and digital meters. We can compare time series images, extract intelligence and draw inference from them. We can count based on movement of objects from and to a defined video frame. And the system can “learn” from all this behavior. This is what we all do right from a few days after birth.

We provides state of art and innovative IoT solutions for different verticals and domains and we have realized the power of “computer vision” in IoT and have done some really interesting experiments and R&D to harness its power. We have also developed hardware that enables this both at the edge level and at the cloud level. The deeper we are going the more we are getting convinced on its power to solve a huge number of digitization use cases. As of now, we are applying its power in security, retail, advertising and many more other areas. If you are interested to know more about it and would like to discuss your requirements with us, please contact us at reachus@smartlinkstechlabs.com with your use case.

Manish Agarwal

B Tech – IIT