In 2012, the VDC Research Embedded Hardware team has been covering several markets including Mil/Aero, and the Embedded Cloud, where we focus on Scalable Edge Nodes (SENs). In this blog, we will look at a several interesting items that stress the need for embedded processing and storage.
Security and surveillance are always a challenge because of the huge increase in the need to capture and archive activities in many areas with fixed cameras. In these cases, there are two main goals. The first is the ability to recognize situations where the appropriate persons need to be alerted and provided with actionable information on what is happening in real time. This is a case where edge processing can provide the needed 24x7 diligence that is difficult or unfeasible to duplicate using human operators. The second goal of these surveillance systems is to provide useful information after the fact. As the resolution and clarity of these surveillance systems improve, they can be used in more cases to identify what happened and the persons involved. The increase in the number of cameras, added resolution, and longer archival periods create huge volumes of data, and, as a result, big opportunities for suppliers.
In Unmanned Aerial Vehicles (UAV), a challenge similar to that of the fixed security applications is seen but, in this case, the UAV is always in motion and the area being observed is variable. Added to that, the UAV is often operating autonomously for its navigation and collision avoidance. This requires massive processing and significant storage in a location where space and operating power are at a minimum. The UAV storage needs to be almost like a DVR with the intelligence to realize that something significant is happening and then alert the operator. The operator can then review the archive as needed. As new CPUs and graphics processors reach the market, embedded computer suppliers are making full use of these in these UAV applications. Although these systems are very capable at present, suppliers like Mercury Computer Systems and GE Intelligent Platforms are being driven to make them even better.
This brings us to yesterday’s Reuters article that highlighted a new satellite imagery system that monitors global forests and detects when illegal activities are taking place. If this works, it will provide authorities with actionable data that will allow them to intercede while the event is taking place. For this to be effective, data from real-time images have to be compared with archived data to determine if changes related to illegal logging, farming, and or mining are taking place. If you consider global covered at the image resolution that is required it drives a tremendous need for storage capacity. When the normal environmental changes in lighting, cloud cover, and growing patterns between image sets have to be accounted for, this represents a huge processing challenge. Together, these requirements would have been financially and/or technically unfeasible just a few years ago. As imaging and processing systems improve, these environmental monitoring applications will become even more powerful and effective.