The BIPcam smart camera from Brain Vision Systems (BVS), Paris, France, emulates how people see things. It models the capabilities of human brains for perception, understanding, and action. It does this through nine real-time data flows that analyze the color, motion, and structure of several objects.
For colors, the camera evaluates luminance, color saturation, and hue, even in low-light settings. The motion-tracking feature reacts instantly to the velocity and direction of moving objects, anticipating their trajectories. Structure recognition includes curves, edges, edge orientation, and magnitudes or size of an edge.
A task manager controls data-flow measurements and statistics and allocates one or more computational tasks to each flow while synchronizing the flows’ operations in the hardware. A multitasking environment lets it track several objects simultaneously.
A CMOS VGA video sensor with a 720 × 480 resolution captures video at 60Œframes/sec, streaming it to a PC via a USB 2.0 interface. An optional UDP/ Ethernet protocol can also be used. With energy needs of only 2.5 W and weight of 2.5 oz. (72 gm), the camera can be incorporated into almost any vision application.
One use of the BIPcam is to analyze traffic conditions for traffic management and control. The camera records the number of vehicles on the road, along with each vehicle’s speed and trajectory, and can adjust traffic signals to create an optimized flow of traffic.
Another automotive use is as a driving assistant, where it tracks colors (such as a dark road versus white lines for line tracking), bends, and inclines in the road.
An open-source software development kit comes with the product for writing source code. The kit also contains several demos, including a minimal video security application.