How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
By Lee J. Nelson
Pedestrians account for more than eleven percent of all domestic roadway fatalities. In general, one-half of those deaths result from pedestrian inattention or error while the other half are the driver’s fault. The US Department of Transportation (DOT; Washington DC) notes that despite some trial countermeasures, no focused effort exists to understand how to apply technology in a coordinated and integrated manner to address this critical safety issue. The DOT’s Federal Highway Administration and the Intelligent Transportation Systems Joint Program Office are exploring further field-testing initiatives for vehicle-based detection and warning systems.
While government agencies ponder the next logical step in promoting safety through technology insertion, automotive industry engineers already are investigating systems that sense impending danger—like a collision or a rollover—and alert the driver accordingly.
Capabilities presently in development merge such logic with computerized control of vehicle braking, steering and suspension components; even sensing a person or an animal in the way. This combination introduces the opportunity to create powerful, automatic collision-avoidance systems with the promise of a dramatic increase in safety on our roadways. Liability concerns and consumers’ desire for improved safety make new protective features a high priority for automakers, both domestically and abroad. Various sensing technologies already underpin applications like adaptive cruise control, parking assist and pre-collision mitigation. Each application includes specialized technology (infrared, LIDAR, RADAR, ultrasonic, etc.) that generally provides either a ranging or a recognition function.
The Road Ahead, Literally
BMW AG (Munich, Germany) recently announced availability for its 7-Series, later this year, of a far-infrared-based night vision system. Night Vision is purported to offer a considerable enrichment of the nighttime driving experience but stops short of attempting to identify image content.
Even given Night Vision’s potential for automatic object identification, the application is very complex: object sensing— especially pedestrian detection and avoidance—requires both ranging and recognition. Combining the two disparate operations can be expensive, difficult to implement and pose important developmental hurdles.