簡易檢索 / 詳目顯示

回結果列表
研究生： 林仕翰
Lin, Shih-Han
論文名稱： 在連續影像中建築物的偵測與定位
Detection and Localization of Buildings in Video Sequences
指導教授： 許秋婷
Hsu, Chiou-Ting
口試委員: 王聖智
孫明廷
許秋婷
學位類別： 碩士
Master
系所名稱： 電機資訊學院 - 資訊工程學系
Computer Science
論文出版年： 2011
畢業學年度： 99
語文別： 英文
論文頁數： 39
中文關鍵詞： 建築物偵測 、建築物定位 、共同建築物偵測
外文關鍵詞： building detection, building localization, co-building detection
相關次數： 點閱：2  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 本論文提出一個在連續影像中偵測與定位建築物的方法。為了準確的找到每張影像畫面(frame)的建築物位置，我們分成兩個階段: 第一階段是在單張影像建立準確可靠的建築機率圖(building probability map)，而第二階段是在連續影像中快速找到共同出現建築物(co-building detection)。在第一階段中我們擷取適合用來描述建築物結構、空間上與紋理性質的特徵(feature)，使用這些特徵資訊去訓練高斯過程分類器(Gaussian Process Classifier)，並使用高斯過程分類器對測試影像分類並建立建築機率圖。當對連續影像的第一張畫面建立好建築機率圖，在第二階段會對此後每一對的連續兩張畫面快速而有效率地找到其中共同出現之建築物。實驗結果顯示，建築機率圖經由共同出現建築物與高斯過程分類器建立的效能相差無幾，但是速度上卻快上了100倍。

    中文摘要 I ABSTRACT II 1. INTRODUCTION 1 2. RELATED WORK 2 2.1 Global feature-based schemes 2 2.1.1 Edge-based features for building detection 2 2.1.2 Multiple features for building detection 3 2.2 Local patch-based schemes 4 2.3 Motivation 5 3. PROPOSED METHOD 10 3.1 Construction of building probability map 10 3.1.1 Feature descriptor for image patches 10 3.1.2 Gaussian process classification 12 3.1.3 Construction of building probability map 14 3.2 Co-building detection from video sequences 14 3.2.1 Frame-based feature distributions 15 3.2.2 Co-building detection 16 4. EXPERIMENTAL RESULTS 24 4.1 Database and setting 24 4.2 Validation of building probability map 25 4.3 Validation of co-building detection 26 4.4 Comparison with other method 27 4.5 Discussion 28 5. CONCLUSION 37 6. REFERENCES 38

    [1] J. Canny, “A Computational Approach to Edge Detection,” IEEE Transacrions on Parrern Analysis and Machine Inrelligence, Vol. 8, pp. 679.698, 1986.

    [2] A. Vailaya, A. K. Jain, and H. J. Zhang, “On Image Classification: City Images vs. Landscapes,” Pattern Recognition, Vol. 31, pp,1921–1936, 1998.

    [3] Q. Iqbal and J. K. Aggarwal, “Applying Perceptual Grouping to Content-Based Image Retrieval: Building images. CVPR, 1999.

    [4] Y. Li and L. G.Shapiro, “Consistent Line Clusters for Building Recognition in CBIR,” In Proceedings of Internetional Conference on Pattern Recognition, 2002.

    [5] J. Malobabic, H. L. Borgne, N. Murphy, N. O'Connor, “Detecting the Presence of Large Buildings in Natural Images,” International Workshop on Content-Based Multimedia Indexing, 2005.

    [6] S. Kumar and M. Hebert, “Man-Made Structure Detection in Natural Images using a Causal Multiscale Random Field,” CVPR, 2003.

    [7] H. Zhou and D. Suter, “Man-Made Structure Segmentation using Gaussian Processes and Wavelet Features,” ICIP, 2007.

    [8] H. Zhou and D. Suter, “Fast Sparse Gaussian Processes Learning for Man-Made Structure Classification,” CVPR, 2007.

    [9] H. T. Chen, “Preattentive Co-Saliency Detection,” ICIP, 2010.

    [10] J. Z. Wang, Integrated Region-Based Image Retrieval: Kluwer Academic Publishers, 2001.

    [11] H.-C. Kim and Z. Ghahramani, “The EM-EP Algorithm for Gaussian Process Classification,” in Workshop on Probabilistic Graphical Models for Classification, 2003.

    [12] C. E. Rasmussen and C. K. I. Williams, Gaussian Processes for Machine Learning: The MIT Press, 2006.

    [13] The LIBSVM tools http://www.csie.ntu.edu.tw/~cjlin/libsvm/

    [14] The GPML tools http://www.gaussianprocess.org/gpml/code/matlab/doc/

    [15] The Man-Made Database http://www.cs.cmu.edu/~skumar/manMadeData.tar

    無法下載圖示 全文公開日期 本全文未授權公開 (校內網路)
    全文公開日期 本全文未授權公開 (校外網路)

    QR CODE