Browse Publications Technical Papers 2007-01-0405

Stereo Vision System for Advanced Vehicle Safety System 2007-01-0405

In this paper, we will introduce a stereo vision system developed as a sensor for a vehicle's front monitor. This system consists of three parts; namely, a stereo camera that collects video images of the forward view of the vehicle, a stereo ECU that processes its output image, and a near-infrared floodlight for illuminating the front at night. We were able to develop an obstacle detection function for the Pre-Crash Safety System and also a traffic lane detection function for a Lane-Keeping Assist System. Especially in regard to the obstacle detection function, we were able to achieve real-time processing of the disparity image calculations that had formerly required long processing times by using two types of recently developed LSIs.
One is a geometry transformation LSI which was exclusively developed for this system to transform left & right camera images into undistorted and parallelized stereo image, another is a parallel processing LSI which was developed to realize a high speed calculation of image processing.
Regarding the stereo vision Pre-Crash Safety System sold in Europe and Japan in the fall of 2006, detection of obstacles, including pedestrians, was made possible through the use of sensor fusion with stereo vision and millimeter wave radar. Support for detecting hazards increased through information regarding distance, speed, width, etc., with a high sensing capability.


Subscribers can view annotate, and download all of SAE's content. Learn More »


Members save up to 18% off list price.
Login to see discount.
Special Offer: Download multiple Technical Papers each year? TechSelect is a cost-effective subscription option to select and download 12-100 full-text Technical Papers per year. Find more information here.
We also recommend:

Wide Field of View (FOV) and High-Resolution Lidar for Advanced Driver Assistance Systems


View Details


The Advanced Sensor Fusion Algorithm for Pre-Crash Safety System


View Details


An Experiment to Non-Intrusively Collect Physiological Parameters towards Driver State Detection


View Details