使用多部攝影機結合電腦視覺技術實現多點觸控之人機互動系統

Multi-Camera Vision-based Finger Detection, Tracking, and Event Identification Techniques for Multi-Touch Sensing and Human Computer Interactive Systems

呂鳴崑(Ming-Kun Lu)

指導教授 ： 陳彥霖

國立臺北科技大學/電資學院/資訊工程系研究所/碩士(2012年)

https://doi.org/10.6841/NTUT.2012.00549

摘要

現今，有許多的多點觸控技術的實現，例如: 電阻式、電容式。然而，由於這些技術的限制，目前不能支援較大的螢幕。因此在本論文將提出並且實作出一套以多點觸控的檢測、追蹤及事件辨識之技術藉以提升應用性。本論文將透過四部紅外線攝影機，捕捉手指觸控到壓克力板的紅外線光，所捕捉到的紅外光點相當於手指的觸控點，可作為輸入設備並且提供便利的人機介面，並結合電腦視覺式的多點觸控技術。與傳統的觸控技術相比，多點觸控技術，允許用戶使用多根手指執行複雜指令。
多點觸控之手指光點偵測，將透過紅外線攝影機取得影像資訊，對影像資訊分離出影像上的明亮點物件，在以多點觸控之手指光點追蹤技術進行追蹤並記錄運動軌跡。然後再利用得到的明亮物件運動軌跡，再進行事件分析，判斷其相對應的事件。在應用方面，希望使用者只要透過簡單的點選動作就可以體驗到較為人性化介面所帶來的樂趣，因此也定義了手指連續點擊和手指滑動軌跡的事件。並且實作出類似桌子形式的觸控螢幕，該裝置可以同時提供多人針對該桌子進行互動。

關鍵字

電腦視覺 ； 多攝影機 ； 多點觸控 ； 人機互動 ； 事件辨識

並列摘要

Nowadays, multi-touch technology has become a popular issue. Multi-touch has been implemented in several ways including resistive type, capacitive type and so on. However, because of limitations, multi-touch by these implementations cannot support large screens. Therefore, this thesis proposes a multi-camera vision-based finger detection, tracking, and event identification techniques for multi-touch sensing with implementation. The proposed system detects the multi-finger pressing on an acrylic board by capturing the infrared light through four infrared cameras. The captured infrared points, which are equivalent to the multi-finger touched points, can be used for input equipments and supply man-computer interface with convenience. Additionally, the proposed system is a multi-touch sensing with computer vision technology. Compared with the conventional touch technology, multi-touch technology allows users to input complex commands.
The proposed multi-touch point detection algorithm identifies the multi-finger touched points by using the bright object segmentation techniques. The extracted bright objects are then tracked, and the trajectories of objects are recorded. Furthermore, the system will analyze the trajectories of objects and identify the corresponding events pre-defined in the proposed system. For applications, this thesis wants to provide a simple human-computer interface with easy operation. Users can access and input commands by touch and move fingers. Besides, the proposed system is implemented with a table-sized screen, which can support multi-user interaction.

並列關鍵字

computer vision technology ； multi-camera ； multi-touch ； human computer interactive ； event identification

參考文獻

[4]Steve Hodges,Shahram Izadi,Alex Butler,Alban Rrustemi,Bill Buxton, "ThinSight: Versatile Multi-touch Sensing for Thin Form-factor Displays."

[7]J. Y. Han, "Low-cost multi-touch sensing through frustrated total internal reflection." in Proc. of UIST 2005, 2005, pp. 115-118.

[8]"A study on image splicing algorithm for large screen multi-touch technique.",2010 International Conference on Machine Vision and Human-machine Interface.

[13]"Open Sound Control." - http://opensoundcontrol.org

[15]Yen-Lin Chen ,Wen-Yew Liang,Chuan-Yen Chiang,Tung-Ju Hsieh,Da-Cheng Lee,Shyan-Ming Yuan,Yang-Lang Chang, "Vision-based finger detection, tracking,and event identification techniques for multi-touch sensing and displaying systems.", Sensors, vol. 11, no. 7, pp. 6868-6892, 2011.