Imagine a camera in the sky that observes a complete city continuously, with enough detail to see the cars driving and the people walking. How can the military and the security agencies utilize this sensor to improve their operations?
There are two main challenges when deploying WAMI systems. In the first place not all video data can be transmitted to the operators on the ground, as the bandwidth from the flying platform is limited. The second problem is that operators cannot process and analyze such large amounts of data. On a normal HD screen only 1% of a WAMI image can be displayed.
A range of smart algorithms has been developed to automatically extract the relevant information from WAMI imagery on-board of the aircraft. This way only the relevant parts of the video are sent to the ground for further analysis by the operator or analyst.
The smart algorithms of TNO include:
- Object detection. People and vehicles are detected using advanced static and dynamic detectors. Our algorithms prevent parallax effects to cause wrong detections.
- Object tracking. People and vehicles are tracked in time using motion models and template matching techniques.
- Tracking repair. Vehicles that are temporarily occluded by buildings, trees or tunnels, are re-identified to ensure long, unbroken tracks.
- Track analytics. Based on the task of the operator, track analytics can be applied to detect certain relevant events. Examples of such events are cars making U-turns or trucks entering in a certain area. Another type of track analytics involves generating patterns of life.
TNO has also developed visualization tools to quickly interact with track sets that include more than 10.000 tracks. These tools can be used by analysts to query about the behaviour of people and vehicles and to immediately playback video of relevant events.
If you are interested in how smart algorithms can increase the effectiveness of WAMI, please contact Rob Kemp for more information.