Position and Vector Detection of Blind Spot motion with the Horn-Schunck Optical Flow

TL;DR

This paper introduces a cost-effective blind-spot detection system using optical flow and vector analysis from live video footage, providing accurate object detection without expensive radar technology.

Contribution

It presents a novel method leveraging optical flow and vector detection for blind-spot monitoring, reducing costs while maintaining accuracy.

Findings

Effective detection of blind-spot objects using optical flow

Reduced system cost compared to radar-based methods

Real-time notification system integrated into vehicle

Abstract

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.