臺灣博碩士論文加值系統

目前道路上架設了許多攝影機但並沒有一個較為完善的管理機制，如欲透過攝影機去追查想找的目標，只能一台一台的調閱資訊，費工費時。有鑑於此，本論文規劃與架構一個有效率的攝影機管理平台，管理相關攝影機之資訊，並且達到追蹤可疑車輛的目的，為了追蹤的結果準確，所以透過了攝影機所回傳的車牌資訊沿著車輛實際行進的路線去進行追蹤，但是這將會遇到一個問題，當搜尋到某一台攝影機有拍攝到該目標車輛時，系統並不知道下一步應該要搜尋哪一台攝影機，所以本文希望在道路地圖上建立邏輯道路攝影機網路拓蹼（Logic-Roadside-Camera Network Topology；LRCNT），以此來確立攝影機與鄰近攝影機之間的關係，如此一來當此系統追蹤車輛的時候只要搜尋到某一台有拍攝到該車輛車牌的攝影機時，只要繼續搜尋該攝影機的鄰近攝影機，再循序往下搜尋最終即可找到該車輛最後可能出現的位置，通報警方到場圍捕。相較以往逐步查找的搜尋方式，本文的方式將較為便捷，並且透過攝影機實際拍攝到的資訊去做追蹤，可達到眼見為憑的效果，而此攝影機管理平台未來也可應用在其他地方。

Currently there are numerous cameras were set up on the road and in general, with no any effective or efficient management mechanism. In other words, if we want to track some specific target through cameras, what can be done and time-expensive is to check through each camera’s information. Therefore, this thesis plans to build a camera management platform to handle the camera-related information, and to achieve the purpose of tracking suspicious vehicles. In order to verify where the tracked results are accurate, the camera shall return vehicles’ license-plate information, along the actual travel route to be tracked. However, when a camera is found to have the shot of the target vehicle, the system does not know which camera should be set for the next search for. As the result, this thesis proposes to create a Logic-Roadside-Camera Network Topology (LRCNT) on the desired road map, in order to establish the relationship between the current and adjacent cameras. Hence, when this system is activated to track some specific vehicle and with a station has captured its license plate, it is simply to continue the searching for the adjacent cameras, and then after the sequentially searching and verification, we can find all the possible positions for the specific vehicle to inform the police departments. Compared to the previous and relative methods, experimental results prove that this proposed approach is more convenient and effective, as well as the captured information can be actually applied to do further tracking to achieve the general goal of seeing is believing. In the coming future, with slight revision, this camera management platform can be applied into the other real-world applications.

中文摘要.........................I
Abstract .....................II
誌謝...........................VI
目錄............................V
表目錄.........................VII
圖目錄....................... VIII
第一章、緒論..................... 1
1.1 前言 .......................1
1.2 文獻回顧 ....................4
1.3 論文架構 ....................6
第二章、建立邏輯道路攝影機網路拓蹼... 7
2.1 邏輯道路攝影機網路拓蹼的定義....... 8
2.2 邏輯道路攝影機網路拓蹼的建立過程... 11
2.3 圖例說明邏輯道路攝影機網路拓蹼........16
第三章、可疑車輛的追蹤機制...... 20
3.1 僅知起點的情形.............24
3.2 僅知範圍的情形.................25
第四章、模擬與實驗結果.............. 27
4.1 地形的比較.....................28
4.2 追蹤過程的效能分析................34
4.2.1實驗樣本地圖....................34
4.2.2比較之方法......................36
4.2.3模擬與實驗結果之總結..............38
第五章、結論........................ 48
參考文獻........................... 49
附錄A：不同地型的比較................ 51

表目錄
表1攝影機權重 ................................................................................................. 22
表2攝影機數的比較 ......................................................................................... 29
表3相同比例尺的情況下不同地圖的比較 ..................................................... 31

圖目錄
圖1.1邏輯道路攝影網路拓蹼建立示意圖 ........................................................ 2
圖2.1 (a)meta leaf (b) leaf ............................................................................... 11
圖2.2 (a)道路取樣圖(b)以某一節點為基準節點 ............................................. 16
圖2.3 (a)探尋鄰近節點 (b)基準節點與其鄰近攝影機建立關聯 .................. 17
圖2.4 (a)以擴展節點轉為基準節點(b)搜尋至無其他擴展節點為止 ............ 18
圖2.5道路攝影機網路拓墣 .............................................................................. 18
圖2.6不同地圖的模擬 ...................................................................................... 19
圖3.1權重模擬圖 .............................................................................................. 22
圖4.1道路地圖 .................................................................................................. 28
圖4.2道路攝影機網路拓蹼 .............................................................................. 29
圖4.3不同地圖的比較 ...................................................................................... 30
圖4.4 (a)道路取樣圖 (b)路口參考點 ............................................................... 32
圖4.5 (a)以某一攝影機為基準 (b)以外圍節點為基準建立葉拓蹼 .............. 32
圖4.6道路路上路口數遠超出攝影機數的情形 .............................................. 33
圖4.7實驗樣本地圖 .......................................................................................... 34
圖4.8已知起點的樣本地圖 .............................................................................. 35
圖4.9攝影機位置圖 .......................................................................................... 35

圖4.10攝影機編號 ............................................................................................ 36
圖4.11 Sequence Search ..................................................................................... 37
圖4.12 Flooding.................................................................................................. 37
圖4.13 LRCNT ................................................................................................... 38
圖4.14攝影機搜尋車牌時的評估時間 ............................................................ 39
圖4.15評估攝影機搜尋時間的地形樣本圖 .................................................... 40
圖4.16攝影機搜尋時間（Sequence Search） ................................................ 40
圖4.17攝影機搜尋時間（Flooding） ............................................................. 41
圖4.18攝影機搜尋時間（LRCNT） ............................................................... 42
圖4.19假設車輛之行進路線（僅知起點） .................................................... 43
圖4.20僅知起點（尋找全程路線） ................................................................ 44
圖4.21僅知起點（找到最後位置） ................................................................ 45
圖4.22僅知範圍的模擬地圖 ............................................................................ 46
圖4.23 假設車輛之行進路線（僅知範圍） ................................................... 46
圖4.24僅知範圍（找到最後位置） ................................................................ 47

[1] Thomas Kuo, Zefeng Ni, Carter De Leo, B.S. Manjunath, “Design and implementation of a wide area, large-scale camera network,” IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2010.
[2] Mahmood Ashoori Lalimi, Sedigheh Ghofrani and Des Mc Lernon, “A Vehicle License Plate Detection Method using Region and Edge Based Methods,” Computers and Electrical Engineering 39, 2013,pp. 834–845.
[3] Amir Sedighi and Mansur Vafadust, “A New and Robust Method for Character Segmentation and Recognition in License Plate Images,” Expert Systems with Applications 38, 2011, pp. 13497–13504.
[4] H.Erdinc Kocer and K.Kursat Cevik, “Artificial Neural Networks Based Vehicle License Plate Recognition,” Procedia Computer Science 3, 2011, pp. 1033-1037.
[5] Minglu Li, Hongzi Zhu, Yanmin Zhu, Lionel M. Ni, “ANTS: Efficient Vehicle Locating Based on Ant Search in Shanghai Grid,” IEEE Transactions on vehicular technology, October, 2009,Vol. 58, No. 8.
[6] T. Ishida and R. Korf, “Moving target search,” in Proc. Int. Joint Conf. Artif. Intell., 1991, pp. 204–210.
[7] V. Bulitko and N. Sturtevant, “State Abstraction for Real-time Moving Target Pursuit:A pilot study,” in Proc. Nat. Conf. Artif. Intell., Workshop Learn. Search, 2006, pp. 72–79.
[8] S. Koenig, M. Likhachev and X. Sun, “Speeding up Moving-Target Search,” in Proc. 6th Int. Joint Conf. Autonom. Agents Multiagent Syst., 2007, article no. 188.
[9] S.Koenig, X. Sun and Y.William, “Generalized Adaptive A*,” in Proc.7th Int. Conf. Autonom. Agents Multiagent Syst., 2008, pp. 469–476.
[10]V. Bulitko and G. Lee, “Learning in Real Time Search: Aunifying Framework,”J. Artif. Intell. Res., vol. 25, 2006, pp. 119–157.
[11]Mingliang Xu, Zhigeng Pan, Hongxing Lu, Yangdong Ye, Pei Lv and Abdennour El Rhalibi, “Moving-Target Pursuit Algorithm Using Improved Tracking Strategy,” IEEE Transactionson Computational Intelligenceand AIin Games, March, 2010, Vol. 2, No. 1.
[12]Q. Lv, P. Cao, E. Cohen, K. Li and S. Shenker, “Search and Replication in Unstructured peer-to-peer Networks,” in Proc. Supercomput., 2002, pp. 84–95.
[13]路網數值圖100年版，交通部運輸研究所發行，內容包含台灣地區之道路、道路節點、行政區界、鐵路運輸、河流、標地物、橋梁等地圖資訊。