How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
By Lee J. Nelson
Canesta, Inc. (Sunnyvale, CA) addresses some of those obstacles by offering both ranging and recognition functions, together in a solitary, small package. The firm’s Electronic Perception Technology functions across multiple, existing, and future applications, and includes novel microchip imagers that are similar in size, complexity and cost to commodity video chips. Unlike conventional video cameras that see the world as flat images, CanestaVision, in realtime, delivers the sensor-to-scene distance for every pixel in an image. Ranging and recognition are furnished by one device whose output can be used for multiple applications aboard the same vehicle.
Canesta maintains that protective features will emerge more rapidly, cost-effectively and be available in a wider selection of vehicle models. Pedestrian detection will become feasible as expensive solutions give way to less pricey, more practical alternatives.
Camera-based vision systems are helping Delphi Delco Electronics (Troy, MI) develop future safety enhancements. Working in partnership with General Motors Corporation (Detroit), the University of Michigan Transportation Research Institute (Ann Arbor, MI) and the DOT’s Volpe Transportation Systems Center (Cambridge, MA), Delphi is re-prioritizing areas where imaging systems can aid motorists by providing extra sets of eyes.
For many applications, one camera captures sufficient image data. Sometimes, though, two cameras are better than one—in a stereo configuration to determine distance.
“Delphi also can combine information from cameras and other types of sensors to enhance system capability,” comments William Shogren, manager of Advanced Vehicle Systems. “For example, RADAR sensors are typically very good at detecting objects around the vehicle, but cameras have the ability to identify objects.” An application might fuse RADAR data—for locating an object—with visible imagery—to classify whether the entity is a pedestrian, a vehicle, a motorcycle, etc. With such classification and location information, the system could take object-specific countermeasures. “If a pedestrian was in danger, perhaps the horn would sound as an alert or a hood airbag would deploy, helping minimize impact,” says Shogren.