• 沒有找到結果。

8.1 結論

本文以混合型演算法進行多目標物體追蹤,根據 7.1 節所驗證之演算法的性 能,我們以粒子群聚最佳化法進行目標物之搜尋,且以粒子濾波器進行目標物之 追蹤。方法為當目標物出現在影像畫面時,追蹤系統以 PSO 來搜尋目標物,而當 其搜尋到目標物後,系統便切換到 PF,對目標物做後續的追蹤,以發揮兩個演算 法之優點,進而提升目標物追蹤的性能。

而硬體實現平台方面,我們利用 SOPC 以軟硬體協同設計之方式,實現混合 型演算法硬體加速的功能,以 NIOS II 軟核心處理器計算權重,並以硬體電路來 更新 PF 與 PSO 之粒子資訊,建立出一個多目標物體追蹤的雛型系統,使得我們 可以以有彈性的方式來設計電路,以達到我們對追蹤系統預期的效能。而當權重 計算模組測試完成後,我們便將其以硬體電路實現出來,進一步提升整體追蹤系 統的效能。

本文以多主從架構為主體,設計影像的擷取電路與顯示電路,並結合軟硬體 協同設計之混合型演算法,實際應用在多目標物體追蹤上。並且因為硬體電路有 平行處理之特性,故當演算法在進行多目標物體追蹤時,依然可以達到即時影像 的效果。

72

8.2 未來展望

由於目前對目標物之建模,是以事先儲存之方式,故未來可利用 LTM 本身 具有的觸碰功能,以點選觸控的方法對目標物進行建模,使得欲追蹤之目標物可 以不必受預先建模之限制。而光源對目標物之顏色比對會有很大的影響,故未來 可加入邊緣偵測、紋理偵測,或利用濾波器處理影像雜訊,以增強影像特徵辨識 的部分。而目標物在移動時,可能會因為尺度變化,而產生追蹤失敗的問題,故 未來可以考慮將 SURF 或 ORB 等演算法應用在硬體追蹤平台上,以解決目標物 尺度變化的問題。

73

參考文獻

[1] Shinya Yamamoto, Yasushi Mae, Yoshiaki Shirai, and Jun Miura, “Realtime multiple object tracking based on optical flows,” Proc. IEEE International conference on Robotics and Automation, Nagoya, Japan, May 21-27, 1995, Vol. 3,

pp. 2328-2333.

[2] 丁士翔,

以時空為基礎之移動物體追尋

,國立中山大學機械與機電工程學系 碩士班碩士論文(指導教授:程啟正),2008。

[3] Valtteri Takala and Matti Pietikainen, “Multi-Object Tracking Using Color, Texture and Motion,” Proc. IEEE Conference on Computer Vision and Pattern Recognition, Minneapolis, MN, June 17-22, 2007, pp. 1-7.

[4] C. Beyan and A. Temizel, “Adaptive mean-shift for automated multi object tracking,” IET Computer Vision, Vol. 6, no. 1, pp. 1-12, January 2012.

[5] Xiaoqin Zhang, Weiming Hu, Wei Li, WeiQu, and Steve Maybank, “Multiple Object Tracking Via Species-Based Particle Swarm Optimization,” IEEE Transactions on Circuits and System for Video Technology, Vol. 20, no. 11, pp.

1590- 1602, Nov. 2010.

[6] J. Sherrah, B. Ristic, and N.J. Redding, “Particle filter to track multiple people for visual surveillance,” IET Computer Vision, Vol. 5, no. 4, pp. 192-200, July 2011.

[7] J. Kennedy and R. Eberhart, “Particle Swarm Optimization,” Proc. IEEE Int. Conf.

Neural Network, Piscataway, 1995, pp. 1942-1948.

[8] C.H. Chen and M.C. Yan, “PSO-Based Multiple People Tracking,”

Communications in Computer and Information Science, Vol. 166, pp. 267-276, 2011.

[9] Chen-Chien Hsu and Guo-Tang Dai, “Multiple Object Tracking Using Particle

74

Swarm Optimization,” 2012 International Conference on Image, Signal and Vision Computing, Oslo, Norway, Aug. 13-14, 2012, pp. 41-44.

[10] Y. C. Ho and R. C. K. Lee, “A Bayesian approach to problems in stochastic estimation and control,” IEEE Transactions on Automatic Control, vol. AC-9, pp.

333-339, 1964.

[11] J. H. Kotecha and P. M. Djuric, “Gaussian Particle Filtering,” IEEE Transactions on Signal Processing, vol. 51, no. 10, pp. 2592-2601, 2003.

[12] N. de Freitas, “Rao-Blackwellised particle filtering for fault diagnosis,” 2002 IEEE Aerospace Conference Proceedings, vol. 4, pp. 4-1767-4-1772.

[13] Y. Jin and F. Mokhtarian, “Variational Particle Filter for Multi-Object Tracking,”

Proc. IEEE 11th International Conference on Computer Vision, Rio de Janeiro, Oct. 14-21, 2007, pp.1-8.

[14] C. Yang, R. Duraiswami, and L. Davias, “Fast multiple object tracking via a hierarchical particle filter,” Proc. Conference on Computer Vision, 2005, vol. 1, pp.212-219.

[15] 林玟玲,

以軟硬體協同設計之混合型即時影像物體追蹤系統

,淡江大學電機 工程學系碩士班碩士論文(指導教授:易志孝),2011。

[16] Altera Corporation, SOPC Builder User Guide, 2003.

[17] Altera Corporation, Designing With Nios & SOPC Builder, 2003。

[18] Shih-An Li, Chen-Chien Hsu, Ching-Chang Wong, and Chia-Jun Yu,

“Hardware/Software Co-design for Particle Swarm Optimization Algorithm,”

submitted to Information Sciences, Aug. 15, 2009. (SCI)

[19] 李世安、翁仲緯、賴鈺婷、翁慶昌,“影像硬體加速器之設計”,2009 中華民 國系統科學與工程研討會,淡江大學,2009 年 6 月。

[20] Chen-Chien Hsu, Shih-An Li, and Wen-Ling Lin “Real-Time Object Tracking

75

Based on Hardware/Software Co-design,” 2011 International conference on Service and Interactive Robotics (SIRCON 2011), Taichung, Taiwan, Dec. 20-21, 2011, pp. 167-170.

[21] Chen-Chien Hsu*, Shih-An Li, and Wen-Ling Lin, “Hardware/Software Co-design of Particle Filter and Its Application in Object Tracking,” Proceedings of the International Conference of System Science and Engineering 2011 (ICSSE 2011), Macao, June 8-10,

2011, pp. 87-91.

[22] 陳郁龍,

改良連續蒙特卡羅法於視訊物件追蹤之研究

,國立東華大學資訊工 程學系碩士班碩士論文(指導教授:許陳鑑),2010。

[23] B. L. Miller and D. E. Goldberg, “Genetic algorithms, tournament selection, and the effects of noise,” Illinois Genetic Algorithms Laboratory Report, no. 95006, July 1995.

[24] 李世安,

即時目標影像追蹤之

SoPC

設計

,淡江大學電機工程學系博士班博 士論文(指導教授:翁慶昌),2008。

[25] 王瑞斌,

以軟硬體協同設計之粒子濾波器及其在影像追蹤之應用

,淡江大學 電機工程學系碩士班碩士論文(指導教授:許陳鑑),2010。

[26] Altera Corporation, Avalon Interface Specifications, Document Version 1.2, 2009.

[27] Altera Corporation, Avalon Interface Specifications, March 2008.

[28] Altera 多 媒 體 發 展 平 台 DE2-70 網 址 , URL:

http://www.altera.com/education/univ/materials/boards/de2-70/unv-de2-70-board [29] Terasic Corporation, URL:http://www.terasic.com.tw/tw/

[30] Terasic Corporation, TRDB_D5M_UserGuide, Document Version 1.0, 2008.

[31] Terasic, THDB-D5M_Hardware specification, Document Version 1.0, 2008.

[32] Terasic, TRDB_LTM_User Manual, Document Version 1.4.1, 2011.

相關文件