I’ve got a new blog post up at EE Times, talking about how to build video devices for the Internet of Things:
Video cameras are uniquely compelling sensors because vision is our dominant sense. Video is invaluable for applications such as license plate recognition, robot navigation, and quality monitoring. Unfortunately, video is also one of the most bandwidth-intensive forms of sensing, so engineers need to carefully consider three areas when designing video systems for the Internet of Things:
- How to partition processing between the camera and the cloud
- What quality of compressed video to upload and at what time
- How to support peer-to-peer video transfers
The canonical IoT model involves a network of sensors collecting data, which is forwarded to cloud servers. The servers analyze the data, often in real-time, to extract pertinent information and initiate appropriate responses. Often no human interaction is required. But because of video’s large inherent data rate, it is important to think carefully about how the necessary data analysis should be partitioned between the camera and the cloud.
Read the whole article here.