English  |  正體中文  |  简体中文  |  全文筆數/總筆數 : 52047/87178 (60%)
造訪人次 : 8682350      線上人數 : 67
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library & TKU Library IR team.
搜尋範圍 查詢小技巧:
  • 您可在西文檢索詞彙前後加上"雙引號",以獲取較精準的檢索結果
  • 若欲以作者姓名搜尋,建議至進階搜尋限定作者欄位,可獲得較完整資料
  • 進階搜尋
    請使用永久網址來引用或連結此文件: http://tkuir.lib.tku.edu.tw:8080/dspace/handle/987654321/88002


    題名: 使用低階攝影機實現機器人視覺式SLAM
    其他題名: Implementation of robot visual SLAM using low-end cameras
    作者: 林冠瑜;Lin, Guan-Yu
    貢獻者: 淡江大學機械與機電工程學系碩士班
    王銀添;Wang, Yin-Tien
    關鍵詞: 特徵偵測;特徵描述;加速強健特徵(SURF);視覺式同時定位與建圖(Visual SLAM);Feature Detection;Feature Description;Speed-Up Robust Features (SURF);Visual SLAM
    日期: 2012
    上傳時間: 2013-04-13 11:56:11 (UTC+8)
    摘要: 低階攝影機有標準化軟硬體介面與價格低的優點,在機器人系統上有普及應用的潛力。缺點是低階攝影機的影像擷取頻率通常低於30Hz,所擷取的影像會有模糊現象。另外,由於即時運算能力的限制,機器人應用視覺感測器通常只能處理低解析度的影像,影像模糊現象很難藉由提高解析度與影像處理方法加以改善。本論文針對所擷取的低解析度與模糊的場景影像,發展有效的影像特徵偵測、描述、與比對方法,建立一致性高的稀疏視覺式特徵地圖,以便應用於機器人同時定位與建圖。在影像特徵描述與比對方面,將規劃強健的特徵描述方法,應用在低解析度的模糊影像,以提高特徵成功比對率。另外,也探討灰階影像區塊特徵描述方法,應用在視覺式特徵地圖的可能性。針對影像特徵偵測與描述方法的效能比較,本研究也規劃特徵可重複性的測試與機器人定位的地面基準測試等實驗。其中特徵可重複性會直接影響視覺式地圖的特徵數量,進而影響狀態估測器的效能。實驗結果顯示本論文所規劃的強健特徵描述方法,在低階攝影機所引起的模糊影像中,能有效提高特徵可重複性與降低地圖的特徵數量。
    Low-end camera has the advantages of standardized hardware and software interfaces with low prices as well as has the potential for a wide application in robotic system. However, the shortcoming is that the image acquisition rate of low-end camera is usually lower than 30Hz. Low image acquisition rate will cause the captured image to be blurred. Due to the limitations of real-time computation, the application of robot vision sensors usually only deals with low-resolution images. Therefore, the methods of enhancing the resolution and advanced image processing will not be considered in robot vision application to resolve the problem of image blurring. In this thesis, an effective image feature detection, description, and matching method will be developed for the low resolution and blurred scene image in order to establish high-persistency visual sparse map in robot simultaneous localization and mapping applications. In feature description and matching procedure, a robust feature description method is planned in this thesis and applied to deal with blurred image and improve the matching rate. Furthermore, the image description method using grayscale image patch will be investigated. The experiments of features repeatability and ground truth are carried out to validate the performance of image feature detection and description. The features repeatability will directly affect the number of feature of the visual map, and further influence the performance of state estimator. The experimental results show that the proposed feature description method can effectively improve the features repeatability in blurred images captured by low-end camera and further reduce the number of map features.
    顯示於類別:[機械與機電工程學系暨研究所] 學位論文

    文件中的檔案:

    檔案 大小格式瀏覽次數
    index.html0KbHTML135檢視/開啟

    在機構典藏中所有的資料項目都受到原著作權保護.

    TAIR相關文章

    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library & TKU Library IR teams. Copyright ©   - 回饋