( 您好!臺灣時間:2021/05/19 09:51
字體大小: 字級放大   字級縮小   預設字形  


研究生(外文):Chih-Min Chen
論文名稱(外文):The Study on Moving Object Segmentation Method for Video Captured by a Moving Camera
指導教授(外文):Thou-Ho Chen、Tsong-Yi Chen
外文關鍵詞:Feature pointMultiple view geometryAffineObject segmentation
  • 被引用被引用:0
  • 點閱點閱:646
  • 評分評分:
  • 下載下載:6
  • 收藏至我的研究室書目清單書目收藏:0
  本文方法可分成三個部分,首先尋找影像中的特徵點(Feature Point),利用多視角幾何(Multiple View Geometry)的原理將特徵點分成前景與背景,第二部分則是找出前景的區域以及利用仿射模型(Affine)找出背景的運動,由仿射模型建立的影像與當前影像的差值,得到移動物的輪廓,最後利用移動物輪廓的移動歷史(Motion History)配合型態學處理(Morphology)切割移動物,並將移動物用矩形框(Bounding Box)框住。
This paper proposed a method for capturing animated objects with a moving camera. Such a method enables moving objects in the video screen, including passengers, motorcycles, and cars, to be segmented and labeled by the cameras themselves. In comparison to fixed cameras, it is relatively difficult to videotape by moving cameras, since in the case of the video filmed by moving cameras, not only do the objects move, but also the frames shift. With the assistance of object segmentation skills, the shapes of the moving objects fail to be effectively segmented and detected. Consequently, in the present study, we utilized a hand-held camera to simulate the condition of a human-shaped robot.
The method proposed in this study can be divided into three parts. We firstly find the Feature Points in the frames, by which, the Feature Points are classified into foreground and background with the assistance of multiple view geometry. Secondly, we find out the zone of the foreground and the movement of background with the assistance of affine. We can get the contour of the motion by comparing the image established by affine and the concurrent one. Finally, we make use of motion history of the continuous motion contour, together with morphology, to circle the moving objects with bounding box.
目 錄
摘 要 i
目 錄 iv
表 目 錄 vii
圖 目 錄 viii
第一章 緒論 1
1.1 研究動機 1
1.2 系統架構與流程 2
1.3 論文架構 3
第二章 相關技術探討 4
2.1 靜態影像之移動物切割: 4
2.1.1背景相減法(Background Subtraction): 5
2.1.2連續影像相減法(Frame Difference): 6
2.1.3光流法(Optical Flow): 7
2.1.3區塊比對法: 8
2.2動態影像之移動物偵測: 9
2.2.1 PTZ攝影機 10
2.2.2 深度與攝影機: 11
2.2.3 單鏡頭攝影機: 13
第三章 相關文獻研究 14
3.1 方向梯度直方圖(HOG) 14
3.2 Identification of Moving Obstacles with Pyramidal Lucas Kanade Optical Flow and k means Clustering 16
3.3 Moving Object Detection by Multi-View Geometric Techniques from a Single Camera Mounted Robot 19
3.4 訓練影像中找出物件的特徵區域 22
第四章 移動攝影機攝取移動物之切割方法 26
4.1 尋找特徵點與分類 28
4.1.1 尋找特徵點 28
4.1.2影像特徵點對應: 30
4.1.3取得前景與背景特徵點: 32
4.2 移動物之區域與邊緣 40
4.2.1 取得前景區域: 40
4.2.2 背景影像重建: 43
4.3 移動物之切割與偵測 47
4.3.1移動物輪廓: 47
4.3.2移動物切割: 48
4.3.3移動物偵測: 52
第五章 實驗結果 54
5.1 系統介面 54
5.2 實驗測試樣本 55
5.3 實驗結果 56
5.4 實驗分析與偵測錯誤 63
5.5 實驗評估 65
第六章 結論與未來工作 70
6.1結論 70
6.2未來工作 71
參考文獻 72
[1]N. Dalal and B. Triggs, “Histograms of Oriented Gradients for Human Detection,” Proc. IEEE International Conference on Computer Vision and pattern Recognition, 2005.
[3]A. Ess, B. Leibe, K. Schindler, Van Gool, Robust Multi-Person Tracking from a Mobile Platform. PAMI 31(10) 2009.
[4]P. Dollar, S. Belongie, and P. Perona. The fastest pedestrian detector in the west. In BMVC, 2010.
[5]A. Ess, B. Leibe, K. Schindler, Van Gool, Improved Multi-Person Tracking with Active Occlusion Handling. 2009.
[6]M. Enzweiler, A. Eigenstetter, B. Schiele, and D. M. Gavrila. Multi-Cue Pedestrian Classification with Partial Occlusion Handling. In CVPR, 2010.
[7]P. Sudowe and B. Leibe. Efficient Use of Geometric Constraints for Sliding- Window Object Detection in Video. In ICVS, 2011
[8]R Hartley and A Zisserman, Multiple View Geometry in Computer Vision, Cambridge University Press, 2004.
[9]W.S.P. Fernando, L. Udawatta and P. Pathirana, Identification of Moving Obstacles with Pyramidal Lucas Kanade Optical Flow and k means Clustering, In 3rd Int’l Conf. on Information and Automation for Sustainability(ICIAfS) (IEEE) , pp. 111-117, Dec. 2007.
[10]S. A. El-Azim, I. Ismail, and H. A. El-Latiff,,”An efficient object tracking technique using block-matching algorithm”, Proc. Of the Nineteenth National, Radio Science Conf.,pp.427-433,2002
[11]L. Di Stefano and E. Viarani. Vehicle detection and tracking using the block matching algorithm. In Proc. of 3rd IMACS/IEEE Int’l Multiconference on Circuits, Systems, Communications and Computer, pages 4491–4496, 1999.
[12]A. Kundu, K M. Krishna and J. Sivaswamy. Moving Object Detection by Multi-View Geometric Techniques from a Single Camera Mounted Robot.IEEE/RSJ International Conference on Intelligent RObots and Systems (IROS), 2009
[13]Abhijit Kundu, C. V.Jawahar and K M. Krishna. Realtime Moving Object Detection from a Freely moving Monocular Camera. To appear in IEEE International Conference on Robotics and Biomimetics (ROBIO), 2010
[14]Q. Zhu, S. Avidan, Y. Mei-Chen and C. Kwang. Fast Human Detection Using a Cascade of Histograms of Oriented Gradients, 2006.
[15]S. Agarwal and D. Roth, “Learning a sparse representation for object detection,” In ECCV’02, pages113-130, 2002.
[16]B. Leibe and B. Schiele, “Interleaved object categorization and segmentation,” In BMVC’03, pages 759-768, 2003.
[17]B. Leibe, A. Leonardis, and B. Schiele, “Combined object categorization and segmentation with an implicit shape model,” In ECCV’04 Workshop on Stat. Learn. in Comp. Vis., pages 17-32, 2004.
[18]B. Leibe and B. Schiele, “Scale invariant object categorization using a scale-adaptive mean-shift search,” In DAGM’04, Springer LNCS, Vol. 3175, pages 145-153, 2004.
[19]B. Leibe, E. Seemann, and B. Schiele, “Pedestrian Detection in Crowded Scenes,” In CVPR’05, pages 878-885, 2005.
[20]L. Wang, J. Shi, G. Song, and I-F. Shen, “Object Detection Combining Recognition and Segmentation,” In ACCV’07, pages 189-199, 2007.
[21]G. Bradski and A. Kaehler “Learning OpenCV Computer Vision with the OpenCV Library” O'Reilly Media, Sep. 2008.
[22]H. Malm, M. Oskarsson, E. Warrant, Adaptive Enhancement and noise reduction in very low light-level video IEEE 11th International Conference on Computer Vision(ICCV),pp 1-8, Oct. 14-21,2007
[23]Y. Wang, C. Sun, M. Chiou, Detection of moving objects in imageplane for robot navigation usingmonocular vision, In EURASIP, 2012.
[24]O. Deniz , G. Bueno, E. Bermejo, R. Sukthankar, Fast and accurate global motion compensation, In CVPR, 2011.
[25]B. Wu and R. Nevatia, Optimizing discrimination-efficiency tradeoff in integrating heterogeneous local features for object detection. In CVPR 2008, pages 1–8, June 2008.
[26]W. Schwartz, A. Kembhavi, D. Harwood, L. S. Davis, Human Detection Using Partial Least Squares Analysis, 2009
第一頁 上一頁 下一頁 最後一頁 top