Position and Vector Detection of Blind Spot motion with the Horn-Schunck Optical Flow
This addresses safety for drivers by providing an incremental improvement in blind spot detection technology.
The paper tackles blind spot detection in vehicles by using optical flow from live image footage to identify objects, notifying the driver with sensory alerts. The result is a cheaper system that avoids expensive radar while maintaining necessary accuracy and adapting to processor speeds.
The proposed method uses live image footage which, based on calculations of pixel motion, decides whether or not an object is in the blind-spot. If found, the driver is notified by a sensory light or noise built into the vehicle's CPU. The new technology incorporates optical vectors and flow fields rather than expensive radar-waves, creating cheaper detection systems that retain the needed accuracy while adapting to the current processor speeds.