WEKO3
アイテム
Moving Targets Tracking On Construction Sites In Complex Scenes
http://hdl.handle.net/10173/1820
http://hdl.handle.net/10173/18205ed2a7d0-46ab-41c5-a66b-34e7a1521c30
| 名前 / ファイル | ライセンス | アクション |
|---|---|---|
SMS10-134.pdf (245.1 kB)
|
|
| Item type | 会議発表論文 / Conference Paper(1) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 公開日 | 2018-02-06 | |||||||||
| タイトル | ||||||||||
| タイトル | Moving Targets Tracking On Construction Sites In Complex Scenes | |||||||||
| 言語 | ||||||||||
| 言語 | eng | |||||||||
| キーワード | ||||||||||
| 主題Scheme | Other | |||||||||
| 主題 | moving targets | |||||||||
| キーワード | ||||||||||
| 主題Scheme | Other | |||||||||
| 主題 | tracking | |||||||||
| 資源タイプ | ||||||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_5794 | |||||||||
| 資源タイプ | conference paper | |||||||||
| 著者 |
Shen, Qiaonan
× Shen, Qiaonan
× An, Xuehui
|
|||||||||
| 抄録 | ||||||||||
| 内容記述タイプ | Abstract | |||||||||
| 内容記述 | This paper mainly discusses the possibility of- and need for-tracking moving targets on construction sites in complex scenes using statically placed video cameras. A reliable target tracking system could work under various conditions, such as changes of lighting and movements of background objects. There are many problems associated with objects detecting and tracking on a construction site, while the major two of them are the changing photometric visual content throughout the course of a day and the presence of moving obstacles. Firstly, to solve the problems for detecting, since the background modeling is one of the most important parts for object detection, this paper presents an effective and adaptive background modeling method for detecting objects on construction sites. The background data is initially obtained from the first several images involving many moving obstacles by improving average value model. Then to deal with the change of lighting, the image is divided into several sub-regions. The background model is updated when the number of different regions between current image and background image is larger than a pre-defined threshold value. Since the scene of construction sites is considered to change slowly, that the updating process does not need to operate at every frame and in every region. As a result, a lot of computational cost is decreasing to ensure real-time object detection in the tracking system. Finally, several video sequences captured from construction sites under different situations are tested and the experimental results compared to the average value model show that the algorithms are efficient and accurate to work under practical construction site conditions. | |||||||||
| 書誌情報 |
Society for Social Management Systems Internet Journal 巻 6, 号 1, 発行日 2010-03 |
|||||||||
| ISSN | ||||||||||
| 収録物識別子タイプ | ISSN | |||||||||
| 収録物識別子 | 2432-552X | |||||||||
| 著者版フラグ | ||||||||||
| 出版タイプ | VoR | |||||||||
| 出版タイプResource | http://purl.org/coar/version/c_970fb48d4fbd8a85 | |||||||||
| 出版者 | ||||||||||
| 出版者 | Society for Social Management Systems | |||||||||
