The Pedestrian Detection Method Using an Extension Background Subtraction about the Driving Safety Support Systems
-
- Muranaka Noriaki
- Faculty of Engineering, Kansai University
-
- Date Kei
- Oki Electric Industry Co., Ltd
-
- Tokumaru Masataka
- Faculty of Engineering, Kansai University
-
- Imanishi Shigeru
- Faculty of Engineering, Kansai University
Bibliographic Information
- Other Title
-
- 安全運転支援システムにおける拡張背景差分を用いた歩行者検出法
- アンゼン ウンテン シエン システム ニ オケル カクチョウ ハイケイ サブン オ モチイタ ホコウシャ ケンシュツホウ
Search this article
Abstract
In recent years, the traffic accident occurs frequently with explosion of traffic density. Therefore, we think that the safe and comfortable transportation system to defend the pedestrian who is the traffic weak is necessary. First, we detect and recognize the pedestrian (the crossing person) by the image processing. Next, we inform all the drivers of the right or left turn that the pedestrian exists by the sound and the image and so on. By prompting a driver to do safe driving in this way, the accident to the pedestrian can decrease.<br>In this paper, we are using a background subtraction method for the movement detection of the movement object. In the background subtraction method, the update method in the background was important, and as for the conventional way, the threshold values of the subtraction processing and background update were identical. That is, the mixing rate of the input image and the background image of the background update was a fixation value, and the fine tuning which corresponded to the environment change of the weather was difficult. Therefore, we propose the update method of the background image that the estimated mistake is difficult to be amplified. We experiment and examines in the comparison about five cases of sunshine, cloudy, evening, rain, sunlight change, except night. This technique can set separately the threshold values of the subtraction processing and background update processing which suited the environmental condition of the weather and so on. Therefore, the fine tuning becomes possible freely in the mixing rate of the input image and the background image of the background update. Because the setting of the parameter which suited an environmental condition becomes important to minimize mistaking percentage, we examine about the setting of a parameter.
Journal
-
- IEEJ Transactions on Electronics, Information and Systems
-
IEEJ Transactions on Electronics, Information and Systems 124 (8), 1575-1584, 2004
The Institute of Electrical Engineers of Japan
- Tweet
Keywords
Details 詳細情報について
-
- CRID
- 1390001204603989632
-
- NII Article ID
- 10013325191
- 40007001325
-
- NII Book ID
- AN10065950
-
- ISSN
- 13488155
- 03854221
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- NDL
- Crossref
- CiNii Articles
-
- Abstract License Flag
- Disallowed