A sub-pixel stereo vision system for cost-effective intelligent vehicle applications

A variety of vision systems for intelligent vehicle applications have utilized multiple cameras. While the performance benefits of these multi-sensor systems are great, the hardware cost of these systems will prove to be challenging, if not prohibitive in production. One avenue to cost reduction is the use of sub-pixel processing techniques, combined with dedicated, mixed processing technology. We have successfully applied sub-pixel processing techniques to a proven multiple camera vision system. This work supports the premise that a sub-pixel vision system will occupy an attractive position on the price/performance curve for intelligent vehicle applications.