簡易檢索 / 詳目顯示

研究生: 許書銘
Hsu, Shu-Ming
論文名稱: 以特徵為基礎之樹葉影像分類系統
A Leaf Image Classification System Based On Image Features
指導教授: 王明習
Wang, Ming-Shi
學位類別: 碩士
Master
系所名稱: 工學院 - 工程科學系
Department of Engineering Science
論文出版年: 2011
畢業學年度: 99
語文別: 中文
論文頁數: 51
中文關鍵詞: 樹葉影像分類支援向量機主成分分析
外文關鍵詞: Leaf Image Classification, Support Vector Machine, Principal Component Analysis
相關次數: 點閱:88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
  • 隨著資訊科技的進步,人們越來越有機會帶著具有照相功能的行動科技產品例如智慧型手機,出外郊遊並且接觸到許多植物,而如何利用手中的科技產品辨別植物的種類,逐漸變成一個受到重視的課題。樹葉是植物中最容易取得的重要成分之一,因此常被用來當作辨識植物種類的主要標的物。本論文提出一個樹葉影像分類系統,以樹葉影像的特徵為分類的基礎,擷取其特徵後建立資料庫,以支援向量機作為分類器,利用主成分分析選取所降低的特徵向量的維度。由實驗結果顯示出本系統所提出的分類方法具有不錯的效果。

    Plant recognition based on features of a leaf image is an attractive issue in the recent years. In this thesis, a leaf image classification system was proposed. For an input leaf image, the image is firstly preprocessed by using the digital image processing algorithm to extract the features as the input vectors of the Support Vector Machine. The Support Vector Machine is trained to classify leaf images. We also use Principal Component Analysis to reduce the dimension of the input vectors. Then we analyses to choose which dimension is proper to train the Support Vector Machine. From the experimental results, it is shown that the proposed method can perform well.

    摘要 i Abstract ii 誌謝 iii 目錄 iv 表目錄 vi 圖目錄 vii 第一章 緒論 1 1.1 研究動機與目的 1 1.2 相關研究 3 1.3 論文架構 6 第二章 背景知識 8 2.1 支援向量機 8 2.1.1 線性可分離支援向量機 9 2.1.2 線性不可分離支援向量機 13 2.1.3 非線性支援向量機 15 2.2 概率類神經網路 16 2.2.1 架構解說 18 2.2.2 學習過程 20 2.2.3 回想過程 20 2.3 主成分分析 21 2.3.1 主成分分析之轉換原理 22 2.3.2 主成分分析的數理特性 22 2.3.3 主成分分析之運算 23 第三章 樹葉影像分類系統 26 3.1 樹葉影像特徵資料庫建立之流程簡介 26 3.2 特徵之提取 30 3.2.1特徵之選取 31 3.2.2樹葉影像彩色特徵之提取 31 3.2.3提取樹葉型態之特徵之前置處理 33 3.2.4提取樹葉型態之特徵 34 3.3 產生樹葉影像之特徵向量 37 3.4 分類系統 38 3.4.1支援向量機分類器 38 3.4.2權重系統 39 第四章 系統實作與實驗結果 41 4.1 系統環境與資料來源 41 4.2 實驗之樹葉資料種類 42 4.3 實驗結果 43 3.3.1選擇特徵維度之測試 43 3.3.2與其他方法之比較 45 4.4 程式實作操作介面與特徵資料範例 42 第五章 結論與未來研究方向 48 5.1 結論 48 5.2 未來展望 49 參考文獻 50

    [1] Yanning Zhang, Rongchun Zhao, “Image Classification by Support Vector Machines,” Proceedings of 2001 International Symposium on Intelligent Multimedia, video and Speech Processing, May 2-4, 2001
    [2] Wai-Tak Wong, Frank Y. Shih, Jung Liu “Shape-based image retrieval using support vector machines, Fourier descriptors and self-organizing maps,” International journal of Information Science(2007)1878-1891
    [3] Qing-KuiMan, Chun-HouZheng, Xiao-Feng Wang and Feng-Yan Lin, “Recognition of Plant Leaves Using Support Vector Machine,” Communications in Computer and Information Science, Volume 15, Part 5, 192-199. 2008
    [4] Krishna Singh, Indra Gupta, Sangeeta Gupta, “SVM-BDT PNN and Fourier Moment Technique for Classification of Leaf Shape,” International Journal of Signal Processing, Image Processing and Pattern Recognition Vol.3, No. 4, December, 2010
    [5] Stephen Gang Wu, Forrest Sheng Bao, Eric You Xu, Yu-Xuan Wang, Yi-Fan Chang and Qiao-Liang Xiang, “A Leaf Recognition Algorithm for Plant Classification Using Probabilistic Neural Network,” IEEE International Symposium on Signal Processing and Information Technology, July 2000
    [6] H. Fu and Z. Chi, “Combined Thresholding and neural network approach for vein pattern extraction from leaf images,” IEEE Proceedings-Vision, Image and Signal Processing, Vol. 153, No. 6, December, 2006
    [7] Krishna Singh, Indra Gupta, Sangeeta, “Plant Species Classification By Leaves Using Neural Network,” International Journal of Electronics and Computers Vol. 2, No. 1, 2009
    [8] Vladimir N. Vapnik, “The Nature of Statistical Learning Theory,” Springer, 1995
    [9] Donald F. Specht, “Probabilistic neural networks,”Neural Networks Vol. 3, No. 1, 1990
    [10] Lindsay I Smith, “A tutorial on Principal Components Analysis.” February 26, 2002
    [11] 葉冠麟, “Principal Component Analysis with Missing Data,” June 16, 2006
    [12] Libsvm,http://www.csie.ntu.edu.tw/~cjlin/libsvm/
    [13] Flavia, A Leaf Recognition Algorithm for Plant Classification using PNN.http://flavia.source.net

    無法下載圖示 校內:2013-09-14公開
    校外:不公開
    電子論文尚未授權公開,紙本請查館藏目錄
    QR CODE