中大機構典藏-NCU Institutional Repository-提供博碩士論文、考古題、期刊論文、研究計畫等下載:Item 987654321/64891
English  |  正體中文  |  简体中文  |  全文笔数/总笔数 : 78728/78728 (100%)
造访人次 : 33549005      在线人数 : 635
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
搜寻范围 查询小技巧:
  • 您可在西文检索词汇前后加上"双引号",以获取较精准的检索结果
  • 若欲以作者姓名搜寻,建议至进阶搜寻限定作者字段,可获得较完整数据
  • 进阶搜寻


    jsp.display-item.identifier=請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/64891


    题名: 利用測繪車影像萃取道路標誌 重建細部道路模型;Using Mobile Mapping System to Extract Road Marking for Road Model Reconstruction
    作者: 施凱倫;Shih,Kai-lun
    贡献者: 土木工程學系
    关键词: 移動式測繪系統;道路標誌辨識;道路模型
    日期: 2014-07-30
    上传时间: 2014-10-15 14:32:55 (UTC+8)
    出版者: 國立中央大學
    摘要: 隨著移動式測繪系統(Mobile Mapping System)的迅速發展,空間資料的蒐集變得更有效率且更有時效性,不論是三維坐標資訊、道路影像序列亦或是光達點雲資料,所需花費的蒐集成本皆大幅降低。因此,如何整合移動式測繪系統及遙測的資料,產生具有屬性資料之細部道路模型為空間資訊領域中的重要課題。本研究是以LOD-2三維道路模型為基礎,利用道路影像配合影像處理與攝影測量技術萃取有興趣的特徵區塊(region of interesting),重點為路面標誌種類辨識演算法的建立,並藉由辨識出的指向線標誌種類,重建具有車道屬性資料之細部道路模型。

    主要的研究內容包含四個部分:(1)特徵區塊萃取,(2)特徵區塊篩選,(3)路面標誌辨識,(4)模型整合。特徵區塊萃取的部分,本研究先利用影像分割的方式從道路影像萃取有興趣的特徵區塊,同時解決少部分標誌被遮蔽的問題。接著,為求取幾何特性,本研究以Harris角點偵測萃取特徵區塊的角點,並利用共線條件式配合移動測繪系統所提供的內、外方位資訊解算各個角點之三維坐標。透過三維坐標決定特徵區塊之幾何特性,再以幾何特性的值篩選之。完成上述步驟後,再利用本研究提出的路面標誌辨識演算法,辦別特徵區塊為何種指向線標誌。最後,將辨識出的指向線標誌擺放至LOD-2道路模型上,重建具有屬性資料的細部道路模型。

    研究範例成果顯示,本研究提出的指向線標誌種類辨識方法,可以有效地判別由道路影像萃取的特徵區塊為何種標誌,將標誌模型置放於道路模型上後,使得道路模型的整體視覺模擬更為擬真。而透過車道屬性資料的建置,能讓使用者在操作的過程中獲取更詳細且更完整的道路資訊。
    ;The technology of Mobile Mapping System (MMS) is advancing rapidly. MMS can acquire image sequences and position data more efficiently and effectively. Thus, the integration of MMS and remote sensing data is an important task in geospatial technologies for various applications, such as road model reconstruction, building up three dimensional attribute databases, and so on. The primary objective of this research is to reconstruct the detailed road model and build up attribute database. Focus is placed on extracting and recognizing road markings, especially direction markings, automatically from MMS images.

    The proposed scheme consists of four major parts: (1) region of interesting (ROI) extraction, (2) ROI filtering, (3) road markings recognition, (4) model integration. In the ROI extraction, the mean-shift based image segmentation is employed for automatically extracting ROI and solving the shadow area resulted from trees or cars. Secondly, in order to extract the features of ROI, Harris corner detector is utilized to detect the vertexes from ROI. Subsequently, three dimensional coordinates of vertexes from ROI are calculated using collinearity condition equations. In the third part, a proposed road markings recognition method is used to recognize ROI with road marking templates. Finally, recognized road markings and attribute information are placed to corresponding locations of Level of detail (LOD) 2 road model based on the calculated coordinates.

    Experimental results demonstrate that ROI extraction is more efficient using the developed image segmentation techniques. Meanwhile the proposed algorithm can improve several ROI features of road markings recognition. In addition, the experimental results prove that using the proposed algorithms, integrating MMS data and other geo-information can reconstruct highly detailed road models with attribute database.
    显示于类别:[土木工程研究所] 博碩士論文

    文件中的档案:

    档案 描述 大小格式浏览次数
    index.html0KbHTML401检视/开启


    在NCUIR中所有的数据项都受到原著作权保护.

    社群 sharing

    ::: Copyright National Central University. | 國立中央大學圖書館版權所有 | 收藏本站 | 設為首頁 | 最佳瀏覽畫面: 1024*768 | 建站日期:8-24-2009 :::
    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - 隱私權政策聲明