以FPGA實現即時手追蹤系統

FPGA Implementation of a Real-Time Hand Tracking System

黎公海(Cong-Hai Le)

指導教授 ： 徐元寶

國立虎尾科技大學/電機資訊學院/資訊工程研究所/碩士(2014年)

https://doi.org/10.6827/NFU.2014.00102

摘要

本文採用一種FPGA平台來實現打磚塊遊戲。在遊戲中，擋板是由手部動作，通過影像處理模组進行控制，手部動作由人體膚色檢測系統來檢測。首先，影像影由CMOS攝影機捕獲，接著經由子樣本模塊來減少影像的複雜度。其次，將RGB格式的輸入影像變換到對數色系座標系統(log -opponent)的顏色空間, 同時經由二進制膚色概率值的選定來過濾膚色像素。第三，本系統利用影像處理分析演算法( image processing analysis algorithm)來處理多皮膚區域的情況。 第四，皮膚區域被檢查以確定是否它們是人類膚色。第五，進行遊戲演算法之計算，並在VGA屏幕上顯示結果。最後，由手部動作控制打磚塊遊戲的擋板。實驗結果顯示，本系統具有良好的性能，並滿足手部動作追蹤的目標。

關鍵字

FPGA ； 圖像處理 ； log-opponent ； 皮膚偵測 ； 打磚塊遊戲

並列摘要

This thesis uses a FPGA-based platform to implement a breakout game system. In the game, the paddle is controlled by hand motions through an image processing module. The system detects the hand motions by human skin color. Firstly, images are captured by a CMOS camera and a sub-sample module is used to reduce the image complexity. Secondly, the RGB format input image is transformed to log-opponent color space. A binary skin probability value is chosen to filter the skin pixels. Thirdly, an image processing analysis algorithm is applied to separate skin regions, if the image has more than one skin regions. Fourthly, the skin regions are checked to determine if they are human skin or not. Fifthly, the game algorithm, brick breakout game, is created and displayed on the VGA screen. Finally, the paddle of the brick breakout game is controlled by hand motions. Designed under hardware and software co-design methodology, the resultant system demonstrates high performance and satisfies the goal of real-time hand motion tracking.

並列關鍵字

FPGA ； image process ； log-opponent ； skin detection ； breakout game

參考文獻

[1] W. Freeman and M. Roth, “Orientation Histograms for Hand Gesture Recognition,” Cambridge, 1994.

[2] H. Kato and T. Kato, “A Marker-Less Augmented Reality Based on Fast Fingertip Detection for Smart Phones,” in International Conference on Consumer Products, 2011.

[3] G. Hackenberg, R. McCall and W. Broll, “Lightweight plam and finger tracking for real-time 3D gesture control,” Singapore, 2011.

[4] A. Malima, E. Ozgur and M. Cetin, “A fast algorithm for vision-based hand gesture recognition for robot control,” Istanbul.

[5] S. Dominguez, T. Keaton and A. Sayed, “A Robust Finger Tracking Method for Multimodal Wearable Computer Interfacing,” 2001.