Summary

We propose an innovative vehicle crash avoidance system that leverages rear lamp detection and backside shade information from vehicles. This system autonomously estimates the distance between vehicles by analyzing the width of the front vehicle in images.

The system initiates the process by identifying potential rear lamps, employing color information such as Hue-Saturation-Value (HSV) and YCbCr. Subsequently, vehicles are detected utilizing a classifier incorporating Haar-like features and the Adaboost algorithm, extracting distinctive characteristics of the vehicle’s rear luminance. If the calculated distance to the front vehicle falls below a predefined threshold, the system issues a timely alert to the driver, preventing an impending collision.

Our experimental results demonstrate the effectiveness of the proposed system under real-world city driving conditions. By integrating advanced image processing techniques and intelligent algorithms, our approach exhibits promising potential in enhancing road safety and mitigating the risks associated with the number of vehicles on the roads.