Abstract

In mobile stereo camera systems, constructing the 3-D surroundings was one of very important processes for safety navigation, but was very time consuming. In order to decrease the number of traffic accidental deaths, pedestrians running out into the street should be efficiently detected. A Pre-Focusing Method (PFM) was proposed for decreasing the time for getting the matching pixel pairs in the stereo pair images. The method PFM provided us pre-focused stereo pair images in which an object at the pre-determined distance had no disparity between the images. The pre-focusing was performed by applying the perspective transformation whose coefficients were obtained by a calibration stage of PFM. An object at the distance was extracted by evaluating the spatial similarity between pre-focused stereo pair images without scanning to find the matching point pairs. For avoiding errors due to edge and uniform area, one of the nonparametric tests, the smooth test, was introduced for evaluating the spatial similarity. A watching window was introduced to detect the pedestrians running out. The method PFM was successfully applied to an actual video data captured by a stereo video camera mounted on a bicycle.

Highlights

  • Various types of autonomous cars are currently being developed and tested on actual roads

  • In the proposed method pre-focusing method (PFM), a perspective transformation is applied to one of the stereo pair images to remove the relative disparity for an object at a pre-determined distance

  • The principle of the proposed method PFM is based on the fact that a quadrangle on a plane in a 3D space is perfectly registered onto any quadrangle on another plane in the 3D space

Read more

Summary

Introduction

Various types of autonomous cars are currently being developed and tested on actual roads. Camera-based systems have been developed, i.e., SUBARU’s Eyesight and HONDA’s Intelligent night vision system(2) These systems commonly acquire threedimensional (3D) environmental road information to control the vehicles without crashing. One is to use a sub-system, such as GPGPU, to accelerate the processing(7), and the other is to adopt a smart method to avoid stereo matching The former is an excellent solution but is very expensive. We propose a pre-focusing method (PFM) to detect an object in a 3D region of interest without the stereo matching process, in local mask scanning to find corresponding point pairs. In the proposed method PFM, a perspective transformation is applied to one of the stereo pair images to remove the relative disparity for an object at a pre-determined distance. To quickly detect pedestrians entering the road, we introduce a watching window to reduce the processing time

Methodology
Perspective transformation
Experimental results and discussion
Extraction of the focused area
Focusing performance
Conclusions
Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call