A Vision-Based Driver Assistance System with Forward Collision and Overtaking Detection

Huei-Yung Lin,Jyun-Min Dai,Lu-Ting Wu,Li-Qi Chen

DOI: https://doi.org/10.3390/s20185139

IF: 3.9

2020-09-09

Sensors

Abstract:One major concern in the development of intelligent vehicles is to improve the driving safety. It is also an essential issue for future autonomous driving and intelligent transportation. In this paper, we present a vision-based system for driving assistance. A front and a rear on-board camera are adopted for visual sensing and environment perception. The purpose is to avoid potential traffic accidents due to forward collision and vehicle overtaking, and assist the drivers or self-driving cars to perform safe lane change operations. The proposed techniques consist of lane change detection, forward collision warning, and overtaking vehicle identification. A new cumulative density function (CDF)-based symmetry verification method is proposed for the detection of front vehicles. The motion cue obtained from optical flow is used for overtaking detection. It is further combined with a convolutional neural network to remove repetitive patterns for more accurate overtaking vehicle identification. Our approach is able to adapt to a variety of highway and urban scenarios under different illumination conditions. The experiments and performance evaluation carried out on real scene images have demonstrated the effectiveness of the proposed techniques.

engineering, electrical & electronic,chemistry, analytical,instruments & instrumentation

What problem does this paper attempt to address?

The paper aims to address several key issues in intelligent vehicle driver assistance systems to improve driving safety and reduce traffic accidents. Specifically, the paper proposes a vision-based driver assistance system that mainly addresses the following three problems:

1. **Lane Change Detection**:
- The system acquires image data through front and rear cameras, detects lane markings, and identifies whether the vehicle is changing lanes. The system uses Region of Interest (ROI) technology, trapezoidal region setting, and an improved Hough transform to detect lane lines.

2. **Forward Collision Warning**:
- By detecting the vehicle ahead and its shadow, the system predicts potential collision risks. The system employs a combination of adaptive ROI and fixed trapezoidal ROI methods to locate the vehicle ahead and proposes a Cumulative Distribution Function (CDF) verification method to improve detection accuracy.

3. **Overtaking Vehicle Identification**:
- Using the rear camera to capture images, the system tracks moving objects through the optical flow method and combines Convolutional Neural Networks (CNN) to remove duplicate patterns, thereby accurately identifying overtaking vehicles.

Overall, this research aims to develop an efficient and robust visual perception algorithm to assist drivers or autonomous vehicles in performing safe lane change operations and preventing forward collision accidents.