系統識別號 | U0002-0706201116325800 |
---|---|
DOI | 10.6846/TKU.2011.00217 |
論文名稱(中文) | 具有主動安全警示與強化動態範圍之立體行車記錄系統 |
論文名稱(英文) | An Active-safety notification stereo car video recording system with enhanced dynamic range |
第三語言論文名稱 | |
校院名稱 | 淡江大學 |
系所名稱(中文) | 資訊工程學系碩士在職專班 |
系所名稱(英文) | Department of Computer Science and Information Engineering |
外國學位學校名稱 | |
外國學位學院名稱 | |
外國學位研究所名稱 | |
學年度 | 99 |
學期 | 2 |
出版年 | 100 |
研究生(中文) | 林宏軒 |
研究生(英文) | Hung-Hsuan Lin |
學號 | 798410113 |
學位類別 | 碩士 |
語言別 | 繁體中文 |
第二語言別 | 英文 |
口試日期 | 2011-06-03 |
論文頁數 | 80頁 |
口試委員 |
指導教授
-
黃連進(micro@mail.tku.edu.tw)
委員 - 黃俊堯 委員 - 陳伯榮 委員 - 黃連進 |
關鍵字(中) |
行車記錄器 增強動態範圍 嵌入式系統 立體相機 |
關鍵字(英) |
Car Video Recorder EDR Embedded Linux Stereo Camera |
第三語言關鍵字 | |
學科別分類 | |
中文摘要 |
行車影像紀錄器的應用範圍越來越廣泛,例如行車紀錄、時間監控、路線監測等等,但是在主動安全上的應用卻不多見,且由於日照及車燈等問題,照成動態範圍過窄,本篇論文希望透過立體影像的加入,不僅可以提供物體距離資訊達到主動安全警示的作用,並利用雙影像輸入的曝光差異透過後處理實現增強動態範圍,進而可以真實還原現場環境環境,本篇論文以嵌入式 Linux平台加上影像壓縮模組配合CMOS感測器以及控制模組,組成一個建置在汽車當中之即時影像記錄器,以此研究作以有多種不同的創新開發,應用在行車安全之中,可以增加駕駛者的生命安全保障,如果不幸發生事故,也藉由記錄事證,可以釐清責任歸屬,平時,也可以藉由違規事證收集舉發進而刺激台灣行車守法的觀念。 本篇論文欲使用兩組含影像擷取、硬體影像壓縮及網路功能之嵌入式系統及後端後置處理服務的伺服器實現立體行車記錄系統。 |
英文摘要 |
The applications of car video recorder have been widely utilized, for example, driving records, timing monitoring, route monitoring, etc. However, there are only few applications of active safety in the current world, and we found issues such as sunshine and car lights, according to a low dynamic range. This paper will through the addition of stereo camera to provide more information of object distance for active safety notifications, and using dual camera different exposure time generate source image through the post-processing to achieve enhanced dynamic range, then the video can recovery to the environment of the scene, this paper will use embedded Linux platform and video compression module, dual camera with CMOS sensor and control module to implement in-car real-time video recorder, use this study as to have a variety of innovative development and application in traffic safety among the can increase the life of driver safety and security of the unfortunate accident, but also by recording things certificate, you can clarify the attribution of responsibility, or, you can also capture illegal driving videos for police.The paper uses two Embedded Linux EV boards, which includes video capture, hardware video compressor, and network interface controller. The system also includes a backend Server to implement stereo car video recording system. |
第三語言摘要 | |
論文目次 |
第一章 前言 1 1.1 研究背景與動機 1 1.2 研究目的 4 1.3 論文架構 4 第二章 文獻回顧 6 2.1 安全煞車距離 6 2.2 Stereo Matching 7 2.3 三角距離量測法 8 2.4 車道障礙物偵測 10 2.5 高動態範圍影像 11 第三章 立體行車影像記錄器 12 3.1 嵌入式系統建置 12 3.2 Stereo Matching 14 3.3 Distance calibration 18 3.4 Record and Upload 23 3.5 ObjDetect 28 第四章 立體行車影像合成與轉換 31 4.1 系統架構 31 4.2 Stereo base EDR 33 4.3 3D 回放服務 39 第五章 研究成果 46 5.1 記錄功能 46 5.2 障礙物警報功能 49 5.3 EDR 53 5.4 YouTube 3D影像 60 第六章 結論與未來研究方向 61 6.1 研究貢獻 61 6.2 未來研究方向 61 參考文獻 63 附錄A. Kernel Configure 67 附錄B. Loaded Module 73 附錄C. Device List 73 附錄D. 英文論文 74 圖目錄 圖1-1 市面上行車記錄器的低動態範圍影像 3 圖2-1 駕駛反應時間及制動時間圖 7 圖2-2 像空間坐標系與物空間坐標系之關係圖 9 圖2-3 三角距離測量法 10 圖2-4 Range of Interest 11 圖3-1 行車影像記錄器架構 13 圖3-2 Embedded Linux 硬體 14 圖3-3 Region of Interest 15 圖3-4 Interest Points 16 圖3-5 Stereo Matching 18 圖3-6 距離實地量測 18 圖3-7 距離量測 19 圖3-8 左右邊不同的可視範圍 20 圖3-9 選取可視範圍 21 圖3-10 可視範圍相等的立體照片 21 圖3-11 對應距離測試標示板選取 22 圖3-12 Recorder架構圖 24 圖3-13 Video Streaming & control Thread流程圖 25 圖3-14 AVI and MP4 file wrapper thread流程圖 26 圖3-15 CloudUploader流程圖 27 圖3-16 ObjDetect 流程圖 29 圖3-17 Match 流程圖 30 圖4-1 系統架構圖 32 圖4-2 動態範圍示意圖 33 圖4-3 高動態範圍示意圖 34 圖4-4 不同曝光值之立體影像 35 圖4-5 亮度柱狀圖 36 圖4-6 過亮範圍 36 圖4-7 過暗範圍 37 圖4-8 深度圖 38 圖4-9a 合成前的圖 39 圖4-9b 合成後的圖 39 圖4-10 左右各自獨立的影片 40 圖4-11 Stereo Video Maker轉換程式 40 圖4-12 Side By Side影片 41 圖4-13 Frame Interlace影片 42 圖4-14 Red Cyan影片 43 圖4-15 YouTube Red Cyan影片 43 圖4-16 Line Interlace影片 44 圖4-17 YouTube Line Interlace影片 45 圖5-1a 連續立體影片截圖 46 圖5-1b 連續立體影片截圖 47 圖5-1c 連續立體影片截圖 47 圖5-1d 連續立體影片截圖 47 圖5-1e 連續立體影片截圖 48 圖5-1f 連續立體影片截圖 48 圖5-1g 連續立體影片截圖 48 圖5-1h 連續立體影片截圖 49 圖5-1i 連續立體影片截圖 49 圖5-2a 70m障礙物測試影像 50 圖5-2b 70m障礙物檢測影像 50 圖5-3a 20m障礙物測試影像 51 圖5-3b 20m障礙物檢測影像 51 圖5-3c 20m障礙物判定圖 51 圖5-4a 2m障礙物測試影像 52 圖5-4b 2m障礙物檢測影像 52 圖5-5a EDR測試圖組5-5 54 圖5-5b 圖組5-5待處理圖及亮度柱狀圖 54 圖5-5c 圖組5-5已處理圖及亮度柱狀圖 55 圖5-5d 圖組5-5細部及其亮度柱狀圖比較 55 圖5-6a EDR測試圖組5-6 56 圖5-6b 圖組5-6待處理圖及亮度柱狀圖 56 圖5-6c 圖組5-6已處理圖及亮度柱狀圖 57 圖5-6d 圖組5-6細部及其亮度柱狀圖比較 57 圖5-7a EDR測試圖組5-7 58 圖5-7b 圖組5-7待處理圖及亮度柱狀圖 58 圖5-7c 圖組5-7已處理圖及亮度柱狀圖 59 圖5-7d 圖組5-7細部及其亮度柱狀圖比較 59 表目錄 表1-1 民國91 至99 年A1 和A2 事故件數、死傷人數 1 表1-2 A1類道路交通事故摘要─按車種別、肇事原因 2 表3-1 Embedded Linux系統 12 表3-2 5x5二階高斯函數係數表 16 表3-3 距離與差距對應表 22 表3-4 限制範圍示意圖 23 表4-1 立體行車影像記錄系統 31 表5-1 障礙物檢測結果 53 表5-2 EDR檢測結果 60 |
參考文獻 |
[1]. M. Bertozzi and A. Broggi , "GOLD: a parallel real-time stereo vision system for generic obstacle and lane detection," Image Processing, IEEE Transactions on, Vol. 7 1, Jan. 1998, pp. [2]. 孫正泰,“以電腦視覺為基礎之道路上障礙物偵測”,碩士論文,國立臺灣大學資訊工程學研究所,台北,2001 [3]. 陳琮仁,“以道路辨識與前車偵測之車載駕駛警示系統”,碩士論文,國立臺灣大學資訊工程學研究所,台北,2003 [4]. 孫瑞鴻,“以電腦視覺為基礎之行車安全輔助系統”,碩士論文,亞洲大學資訊工程學系,台中,2008 [5]. 林益安,“機械視覺應用於不同深度與高度之3D物件機械手臂之自動夾取”,碩士論文,彰化師範大學工業教育學系,彰化, 1999. [6]. 王豐良,“應用 GPS 與非量測型相機於行動測繪車之研究”,碩士論文,國防大學中正理工學院軍事工程研究所,桃園大溪,五月2008. [7]. 黃文利,“近景攝影測量應用於三維建物模型側面影像敷貼之研究”,碩士論文,國立成功大學航空測量研究所,台南, 2001。 [8]. 龔威儒,“局部遮蔽圖形之自動比對”,碩士論文,國立清華大學動力機械工程學系,新竹,2004 [9]. 傅秉綱,“三維建物模型表面影像敷貼自動化之研究”,碩士論文,國立成功大學航空測量研究所,台南, 2002。 [10]. 張維哲,“立體顯影之高動態範圍影像合成”,碩士論文,國立中正大學電機工程所,嘉義, 2008。 [11]. Paul E. Debevec and Jitendra Malik, ”Recovering High Dynamic Range Radiance Maps from Photographs”, Berkeley, In SIGGRAPH 97, Aug. 1997 [12]. Paul E. Debevec “Rendering Synthetic Objects into Real Scenes:Bridging Traditional and Image-based Graphics with Global Illumination and High Dynamic Range Photography”, Berkeley, In SIGGRAPH 98, Aug. 1998 [13]. 丁榮豐,“基於區塊的亮度特性及熵值合併多張不同曝光值影像之高動態範圍影像生成法”,碩士論文,國立成功大學電腦與通信工程研究所,台南,2009 [14]. 林家均,“基於環狀取樣以向量為特徵的新式比對技術”,碩士論文,亞洲大學生物資訊學系,台中,2009 [15]. 林進雄,“行車安全貼身護衛”,碩士論文,義守大學電子工程研究所,高雄,2005 [16]. 楊淑娟,“使用數位式行車紀錄器之駕駛安全研究”,碩士論文,逢甲大學交通工程與管理所,台中,2006 [17]. 交通部, “機動車輛登記數”,交通統計要覽,http://www.motc.gov.tw,2011 [18]. 內政部警政署, “道路交通事故(A1及A2類)”,100年警政統計資料,http://www.npa.gov.tw/,2011 [19]. 內政部警政署, “97年全國道路交通事故概述及交通執法績效分析”,100年警政統計資料,http://www.npa.gov.tw/,2011 [20]. Jian Sun, Nan-Ning Zheng and Heung-Yeung Shum , " Stereo Matching Using Belief Propagation," Image Processing, IEEE Transactions on, Vol. 25 No. 7, Jul. 2003, pp [21]. Brandon M. Smith, Li Zhang, Hailin Jin, “Stereo Matching with Nonparametric Smoothness Priors in Feature Space,” IEEE C, June 2009 [22]. Payman Moallem, Karim Faez, Javad Haddadnid “Fast Edge-Based Stereo Matching Algorithms through Search Space Reduction,” IEIEC, Nov. 2002 [23]. Payman Moallem, Karim Faez, Javad Haddadnid “Search Space Reduction in the Edge Based Stereo Correspondence,” Electrical Engineering Department, Amirkabir University of Technology, 2002,pp [24]. Qingxiong Yang, Liang Wang, Ruigang Yang, Shengnan Wang, Miao Liao1 David Nister , “Real-time Global Stereo Matching Using Hierarchical Belief Propagation,” Center for Visualization and Virtual Environments, Department of Computer Science, University of Kentucky and Department of Electronic Science and Technology, University of Science & Technology of China, Hefei, China, 2006 [25]. Google YouTube, “YouTube 3D mode”, YouTube 3D mode selector, http://www.YouTube.com/select_3d_mode, 2011 [26]. D. G. Lowe, “Distinctive image features from scale-invariant keypoints,” International Journal of Computer Vision, Vol. 60, No. 2, pp. 91-110, 2004. |
論文全文使用權限 |
如有問題,歡迎洽詢!
圖書館數位資訊組 (02)2621-5656 轉 2487 或 來信