透過您的圖書館登入
IP:44.195.47.227
  • 學位論文

鋪面剖面掃描儀應用於路面標線完整度之辨識

Pavement Markings Detection and Identification Using Pavement Profile Scanner

指導教授 : 周家蓓
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


在鋪面自動化檢測中,關於路面標線的研究大多著重在車道線的偵測,並且對於標線的破壞程度並沒有太多深入的探討。路面標線雖然不與鋪面服務績效有直接相關,卻仍然與用路人的行車安全息息相關,特別是路段中央的標字與指向線更是提供了用路人許多必要的資訊,因此標線的定期檢測與維持其功用是相當重要的。一般標線的繪製與維護大多在道路新工或整建之後,然而這樣子的維護頻率,並無法滿足某些交通繁忙路段的維修需求,而以人工檢測的方式來決定重繪與否,不只難以得到客觀的結果,更十分仰賴人力,由此可知,對於路面標線檢測的自動化是必要的。國立台灣大學土木工程所交通工程組引進一台PPS-2005 雷射鋪面掃描儀,此儀器可以獲得路面剖面以及相應的雷射回波強度,其中回波強度可以反映出路面不同材質的差異,而鋪面與標線的更是可以透過回波強度差異來分辨。由於PPS檢測得的資料並不像傳統的影像資料需要透過投影校正,並且不需要仰賴光線照明,在資料的準確性有其優勢,因此相當適合發展標線的自動檢測。 本研究對於橫向線段、行人穿越道、車道分隔線、標字以及指向線分別提出辨識邏輯,透過投影量分析法找出各種標線的位置,辨別其種類,並計算其完整度。在一般標線部分,對於行人穿越道提出了行人穿越道遮罩,以增加行人穿越道與橫向線段的投影量差異,成功的搜尋到各種行人穿越道的位置。對於標字以及指向線,則提出樣板的製作方法以及比對方式,對於完整及破損的標字都有良好的辨識效果。

並列摘要


The studies on the pavement marking in automatic detection of the pavement tend to focus on the detection of the lane line, but the extent of damage of the markings. Although the road markings is not related to the pavement service performance, it is still closely related with the safety of the drivers. The markings on the central section of the road also provide the drivers many of the necessary information. Therefore, the regular examination and maintaining the function of the markings are very important. The painting and maintaining of the road markings are usually done after a road is constructed or resurfaced. However such a maintaining frequency of markings can’t meet some of the maintenance needs on heavy traffic roads. The way to determine whether or not to repaint the markings by manual inspection can be not only difficult to obtain objective results, but also dependent on manpower. From this, automatic detection of road markings is necessary. Pavement Profile Scanner(PPS) can obtain the pavement profile and the related laser reflectivity which has the difference between different materials, and road markings and pavement can be distinguished by the laser reflectivity. Since the information obtained by PPS is unlike the conventional image data that requires the projective correction, and doesn’t rely on illumination, the information has its advantages on the accuracy. It is suitable for the development of the automatic detection of road markings. The transverse lines, crosswalks, lane dividers, character markings and direction markings are separately proposed identification logics, including finding the locations of markings through projective quantity analysis, identifying the types of markings, and calculating their completeness. The crosswalk mask is proposed to increase the difference between the projective quantity of crosswalks and transverse lines, and it is helpful in finding the locations of crosswalks. The template making prodedure and the matching method between samples and templates are proposed for the recognition of character markings and direction markings, and they are both useful in indentifying either a complete marking or a damaged one.

參考文獻


[7] 丘宜謙,市區道路標線抗滑性能之研究,國立台灣大學土木工程學系交通工程組碩士論文,民國95年10月
[18] 李柏,鋪面剖面掃描儀應用於鋪面狀況辨識與偵測之研究,國立台灣大學土木工程學系交通工程組碩士論文,民國101年7月
[17] 蔡鎮宇,鋪面雷射掃描儀應用於鋪面績效分析,國立台灣大學土木工程學系交通工程組博士論文,民國100年7月
[16] 曾逸鴻,中文文件影像中之特殊字體偵測,科學與工程技術期刊,第三卷 第四期,pp29-39,民國九十六年
[11] 李杰儒,智慧型執法系統平台之研究以道路環境辨識演法為基礎,淡江大學運輸管理學系運輸科學碩士班碩士論文,民國97年1月

被引用紀錄


曾奕鈞(2015)。道路標字標線之自動化辨識與重繪標準探討〔碩士論文,國立臺灣大學〕。華藝線上圖書館。https://doi.org/10.6342%2fNTU.2015.01289

延伸閱讀