舞台人物自動追蹤燈系統之研發

Development of Automatic Tracking System of Followspot for Characters on Stage

林振翔(Chen-Hsiang Lin)

指導教授 ： 周瑞仁

國立臺灣大學/生物資源暨農學院/生物產業機電工程學研究所/碩士(2005年)

https://doi.org/10.6342/NTU.2005.00523

摘要

本研究發展一套成本低廉，操作容易之舞台人物自動追蹤燈系統。此系統利用機器視覺方式擷取舞台影像，並以此資訊偵測目標之位置，自動將舞台燈光投射於舞台上的目標人物上，並持續追蹤與照射。它以CCD與個人電腦擷取影像並經影像處理及追蹤演算法計算後，透過研華公司生產之PC-LabCard 818H介面卡輸出脈衝訊號，驅動步進馬達自動操控追蹤燈機構。
本研究同時提出一套Adaptive Mean Shift影像色差追蹤演算法。其計算相鄰時間序列影像之UV差值以求得舞台移動中物體之動態資訊，並進一步結合Kalman Filter 與Mean Shift演算法預測、修正影像中目標人物所在位置。其中Mean Shift演算法之搜尋半徑乃運用Kalman Filter預測目標人物之速度作動態調整，以提升演算法之搜尋效率與正確性。此外，追蹤演算法將擷取影像之RGB值轉換至CIE LUV色彩空間後，僅針對其中代表色度資訊之U、V值計算影像色差，可減少舞台影像因照明改變所造成之追蹤誤差。

關鍵字

機器視覺 ； 卡爾曼濾波器 ； 移動平均 ； 追蹤燈

並列摘要

The study developed a cost-effective and easy-operating automatic followspot system lighting upon a specified character on stage. The machine vision system was designed to automatically acquire images by a CCD camera, then postion and track the target and finally output control signals to the followspot system to aim at the target through Advantech PC-LabCard 818H.
The study also proposed a tracking algorithm combining Kalman Filter and Mean Shift Algorithm to enhance the prediction and identification of the character position in an image. The Adaptive Mean Shift considered color differences as a weightings, obtained from the subtraction between two successive images. And, the searching radius of Mean Shift is adjusted by the state prediction of Kalman Filter to improve the efficiency and accuracy of tracking. In addition to reduce tracking error caused by various illumination conditions on stage, we considered only U and V values in CIE LUV color space to calculate the color differences in successive images.

並列關鍵字

Machine vision ； Kalman filter ； Mean shift ； Followspot

參考文獻

5. Baumberg, A. and D. C. Hogg. 1994. An efficient method for contour tracking using active shape models. In “Proc. of IEEE Workshop on Motion of Non-rigid and Articulated Objects”,194-199.

7. Cheng, Y. 1995. Mean shift, mode seeking, and clustering. IEEE Transactions on Pattern Analysis and Machine Intelligence 17(8):790-799.

8. Comaniciu, D., V. Ramesh and P. Meer. 2000. Real-time tracking of non-rigid objects using mean shift. In “Proc. IEEE Conference on Vision and Pattern Recognition”, 2(2):142-149.

9. Ellis, T. 2002. Multi-camera video surveillance. In “Proc. of the IEEE 36th Annual 2002 International Carnahan Conference on Security Technology.”, 228-233.

10. El-Nasr, M. S. and I. Horswill. 2004. Music and Games: Automating Lighting Design for Interactive Entertainment. Computers in Entertainment (CIE), 2(2).