隨著都市化快速發展，交通量與停車需求持續攀升，停車場經常出現車位難尋、動線壅塞及重複繞行等問題，不僅降低駕駛的停車體驗，也造成額外能源消耗與管理負擔，尤其在地下或室內等封閉型態的停車場由於難以依賴 GPS 進行精準定位與路徑導引，傳統僅根據剩餘車位數量或單點感測器資訊的引導方式，往往無法同時兼顧「即時掌握車輛位置」與「依據整體車流動態調整動線」。為改善上述情形，本研究整合多攝影機車輛追蹤與動態路徑規劃技術，設計並實作一套智慧停車引導原型系統。
本研究以縮尺實體模型停車場為實驗場域，建立環狀道路搭配集中式停車區之配置。系統前端建構多攝影機視覺感知模組，採用深度學習模型 YOLOv11 進行車輛與車牌偵測，結合 BoT-SORT 演算法達成多目標車輛追蹤，並利用 PaddleOCR 進行車牌文字辨識以賦予車輛唯一身分，透過事先定義之道路區域與車位對應關係，將跨攝影機的偵測結果整合為單一且連續的車輛軌跡與車位占用資訊。系統後端則建立圖形化路徑規劃模型，以改良式 A* 演算法為核心，依據當前車流分布、候選車位狀態與動線通行條件，為各車輛指派合適車位並規劃停車路徑，當偵測到道路封閉或局部壅塞時，能即時啟動動態路徑重規劃，以調整各車輛之行進路徑，規劃結果進一步以地圖與導引資訊的形式視覺化呈現，模擬管控端監控介面以及入口與各路口之導引電子看板，提供駕駛清楚的行經指引與轉向提示。
實驗結果顯示，本系統能在多輛車同時入場的情況下，維持車輛跨攝影機的連續追蹤，並即時完成可用車位之分配與路徑指引；於部分道路封閉或局部壅塞情境下，亦能透過改良式 A* 演算法成功重新規劃替代路徑，減少不必要的繞行與等待時間。綜合而言，實驗驗證了結合多攝影機車輛追蹤與動態路徑重規劃應用於智慧停車引導之可行性，所提出之原型系統可作為未來實際停車場導入智慧化管理機制之設計參考。

With the rapid development of urbanization, traffic volumes and parking demand continue to increase, and parking facilities frequently suffer from difficulties in finding vacant spaces, circulation congestion, and repeated circling. These issues not only degrade drivers’ parking experience, but also lead to additional energy consumption and management burden. In particular, in enclosed environments such as underground or indoor parking garages, it is difficult to rely on GPS for accurate localization and route guidance. Conventional guidance approaches that depend solely on the number of remaining parking spaces or single-point sensor information often fail to simultaneously provide real-time awareness of vehicle locations and dynamically adjust traffic flow according to overall congestion conditions. To address these problems, this study integrates multi-camera vehicle tracking and dynamic path planning techniques, and designs and implements a prototype intelligent parking guidance system.
A scaled physical parking-lot model is adopted as the experimental testbed, featuring a ring-shaped circulation lane and a centralized parking area. On the front-end side, a multi-camera vision perception module is constructed. YOLOv11, a deep learning–based object detection model, is employed for vehicle and license-plate detection, while the BoT-SORT algorithm is utilized to achieve multi-object vehicle tracking. PaddleOCR is further applied to recognize license-plate characters and assign a unique identity to each vehicle. Based on predefined mappings between roadway regions and parking spaces, detection results from different cameras are fused into a unified and continuous representation of vehicle trajectories and parking-space occupancy. On the back-end, a graph-based path-planning model is established with an improved A* algorithm as the core. The algorithm assigns an appropriate parking space to each vehicle and plans its parking route according to the current traffic distribution, candidate space status, and lane accessibility. When road closures or local congestion are detected, dynamic path replanning is triggered in real time to adjust vehicle routes. The planning results are further visualized as maps and guidance information, emulating a control-center monitoring interface and guidance displays at the entrance and intersections, thereby providing drivers with clear routing instructions and turning indications.
Experimental results show that the proposed system can maintain continuous cross-camera tracking of vehicles and perform real-time allocation of available parking spaces and route guidance, even when multiple vehicles enter the parking lot simultaneously. Under scenarios involving partial road closures or localized congestion, the improved A* algorithm can successfully replan alternative routes, thereby reducing unnecessary circling and waiting time. Overall, the experiments demonstrate the feasibility of integrating multi-camera vehicle tracking and dynamic path replanning into intelligent parking guidance, and the proposed prototype system can serve as a design reference for future deployment of intelligent management mechanisms in real-world parking facilities.

目錄
第一章	緒論	1
1.1	研究背景與動機	1
1.2	研究目的	4
1.3	論文架構	6
第二章	背景技術與相關研究	8
2.1	多攝影機車輛偵測與追蹤技術概述	8
2.1.1	物件偵測	9
2.1.2	物件追蹤	11
2.1.3	文字辨識	13
2.1.4	影像串流	15
2.2	路徑規劃演算法	17
2.2.1	停車場環境建模技術	17
2.2.2	路徑規劃演算法與A*演算法概述	18
2.3	系統決策與整合技術	20
2.3.1	有限狀態機	20
2.3.2	異質系統資料交換與非同步通訊技術	21
2.3.3	文獻總結與研究切入點	22
第三章	系統架構	24
3.1	系統總體架構	24
3.2	硬體設備介紹	25
3.3	分散式影像串流節點配置	28
3.4	實驗環境	30
第四章	導引系統設計與實作	32
4.1	導引系統運作流程	34
4.2	前端影像擷取與預處理	34
4.2.1	多攝影機與邊緣運算裝置配置	34
4.2.2	影像串流程式設計與即時性考量	35
4.2.3	影像預處理與有效區域設定	36
4.2.4	前端影像資料與後續模組之銜接	38
4.3	車輛偵測、追蹤與身份建立	38
4.3.1	車輛偵測	38
4.3.2	多目標追蹤	38
4.3.3	車牌文字辨識與車輛身分建立	39
4.3.4	區域定位與進出場事件產生	41
4.3.5	跨攝影機整合與交遞機制	42
4.3.6	停車格判定區域設計與停車行為辨識方法	44
4.3.7	系統狀態監控與實驗輔助介面設計	48
4.4	狀態建模與資料交換	50
4.4.1	導引模組狀態資料定義	51
4.4.2	狀態同步與更新時機	51
4.4.3	事件流資料格式	52
4.4.4	導引結果輸出格式	53
4.5	停車路徑規劃與動態重規劃	54
4.5.1	停車場路網建模	54
4.5.2	A* 成本函數與啟發式設計	56
4.5.3	改良式 A* 之動態成本設計與封閉處理	58
4.5.4	重規劃觸發條件與抖動避免機制	59
4.5.5	停車行為造成之暫時性道路堵塞處理	60
4.5.6	規劃結果轉換為導引指令之映射方式	62
4.6	導引結果顯示與前端介面設計	63
第五章	系統實作結果與展示	66
5.1	實驗場域設定	66
5.2	停車導引實驗	67
5.2.1	車輛入場偵測、路徑規劃與電子看板導引展示	67
5.2.2	多車輛導引與動態重規劃路徑	72
5.2.3	無可行路徑之特殊情形	88
5.2.4	後續導引流程展示	108
第六章	結論與未來展望	129
6.1	結論	129
6.2	未來展望	131
參考文獻	133

圖目錄
圖 1 導引系統整體架構圖	24
圖 2 NVIDIA Jetson Orin Nano Super	26
圖 3 停車場實驗場域佈局與配置示意圖	30
圖 4 導引系統運作流程圖	33
圖 5 停車場車道區域判定劃分與車輛辨識框示意圖[16]	36
圖 6 影像座標輔助白線示意圖（以 Road A North 攝影機畫面為例）	37
圖 7 停車場道路區域劃分示意圖	41
圖 8 SLOT 判定區域顯示	45
圖 9 系統操作與狀態監控介面	49
圖 10 停車場即時狀態地圖	55
圖 11 導引電子看板架設位置示意圖	65
圖 12 模擬各路之導引電子看板之前端介面	65
圖 13 實際實驗場域圖	67
圖 14 偵測模組視覺化GUI介面	68
圖 15 車輛偵測框與標記資訊	69
圖 16 即時狀態更新示意圖	69
圖 17 文字導引介面示意圖	70
圖 18 入口處電子看板顯示導引指示示意圖	72
圖 19 車輛A入場	73
圖 20 車輛A於入口處	74
圖 21 車輛A之即時路徑文字導引	74
圖 22 電子看板導引車輛A行駛方向	75
圖 23 車輛B入場	76
圖 24 車輛A與車輛B即時位置	76
圖 25 車輛A與車輛B之即時路徑文字導引	77
圖 26 入口處電子看板導引車輛B行駛方向	77
圖 27 車輛A駛入Road D Midsection目標停車位NW前	78
圖 28 車輛A位於 Road D Midsection即時狀態顯示	78
圖 29 車輛A轉向準備停車	79
圖 30 車輛A正在停車導致Road D 封閉	80
圖 31 車輛B改變停車路徑	81
圖 32 入口處電子看板顯示新導引給車輛B	82
圖 33 車輛B被攝影機偵測駛入Road A Midsection區域	83
圖 34 車輛B行駛至Road A Midsection區域	83
圖 35 電子看板Road A South顯示對於車輛B之導引	84
圖 36 車輛B被攝影機偵測駛入Road B Midsection區域	85
圖 37 車輛B行駛至Road B Midsection區域	85
圖 38 電子看板Road B East顯示對於車輛B之導引	86
圖 39 車輛B被攝影機偵測駛入Road C Midsection區域	87
圖 40 車輛B行駛至Road C Midsection區域	87
圖 41 電子看板Road C North顯示對於車輛B之導引	88
圖 42 車輛C入場	89
圖 43 車輛C於入口處位置	90
圖 44 車輛C之即時路徑文字導引	91
圖 45 入口處電子看板導引車輛C行駛方向	91
圖 46 車輛C被攝影機偵測駛入Road A Midsection區域	92
圖 47 車輛C行駛至Road A Midsection區域	92
圖 48 電子看板Road A South顯示對於車輛C之導引	93
圖 49 車輛C駛入Road B Midsection目標停車位SW前	94
圖 50 車輛C於目標停車位SW前即時狀態顯示	94
圖 51 車輛D入場	95
圖 52 車輛D於入口處	95
圖 53 車輛D即時路徑文字導引	96
圖 54 入口處電子看板顯示新導引給車輛D	97
圖 55 車輛C正在停車	98
圖 56 車輛C正在停車導致Road B暫時封閉不可通行	99
圖 57 因車輛C正在停車導致道路封閉觸發停車路徑重規劃	100
圖 58 電子看板因車輛C正在停車而更新新導引	101
圖 59 車輛D被攝影機偵測駛入Road A Midsection區域	102
圖 60 車輛D正在行駛至封閉路段Road B前	102
圖 61 電子看板Road A South顯示對於車輛D之導引	103
圖 62 車輛A完成停車	104
圖 63 車輛A完成停車並標示符號X表示	105
圖 64 因路段Road D封閉狀態排除而重規劃路徑	106
圖 65 電子看板Road A South導引車輛D迴轉行駛	107
圖 66 電子看板Road A North 接續導引車輛D	108
圖 67 車輛D被攝影機偵測駛入Road D Midsection區域	109
圖 68 車輛D駛入Road D Midsection區域即時狀態顯示	109
圖 69 電子看板Road D East顯示對於車輛D之導引	110
圖 70 車輛B與車輛D會車	111
圖 71 車輛B與車輛D會車之即時狀態顯示	111
圖 72 車輛D駛入Road C Midsection區域且車輛B準備停車	112
圖 73 車輛D行駛至Road C Midsection區域即時狀態顯示	113
圖 74 電子看板Road C South顯示對於車輛D之導引	113
圖 75 車輛B正在停車	114
圖 76 車輛B正在停車導致Road D封閉	115
圖 77 因Road D發生封閉狀態而重規劃路徑	116
圖 78 車輛B顯示於正在停車欄位	116
圖 79 車輛C完成停車	117
圖 80 車輛C完成停車並標示符號X表示	118
圖 81 Road B封閉排除後重規劃之路徑	119
圖 82 車輛C完成停車後被清除正在停車訊息	119
圖 83 車輛D行駛至目標停車位SE前	120
圖 84 車輛D行駛至目標停車位SE前準備停車	120
圖 85 車輛B完成停車	121
圖 86 車輛B完成停車並標示符號X表示	122
圖 87 Road D封閉排除後重規劃之路徑	123
圖 88 車輛B完成停車後被清除正在停車訊息	123
圖 89 車輛D正在停車	124
圖 90 車輛D正在停車導致Road B暫時封閉	124
圖 91 車輛D顯示於正在停車欄位	125
圖 92 車輛D完成停車	126
圖 93 車輛D完成停車後即時狀態顯示	127
圖 94 入場車輛皆停車完畢	128
圖 95 電子看板因車輛全數停車完畢而無新指引	128

表目錄
表 1 攝影機規格	27

[1]	X. Sui, X. Ye, T. Wang, X. Yan, J. Chen, and B. Ran, "Microscopic simulating the impact of cruising for parking on traffic efficiency and emission with parking-and-visit test data," International Journal of Environmental Research and Public Health, vol. 19, no. 15, pp. 9127, 2022. 
[2]	Y. Zhu, X. Ye, J. Chen, X. Yan, and T. Wang, "Impact of cruising for parking on travel time of traffic flow," Sustainability, vol. 12, no. 8, pp. 3079, 2020. 
[3]	W.-H. Tsai, H.-C. Chen, K.-H. Lee, and C.-F. Yang, "Implementation of an AIoT-based Smart Parking System for Urban Mobility and Sustainable Infrastructure Management," Sensors & materials, vol. 37, 2025. 
[4]	S. Saki and T. Hagen, "Cruising for parking again: Measuring the ground truth and using survival analysis to reveal the determinants of the duration," Transportation Research Part A: Policy and Practice, vol. 183, pp. 104045, 2024. 
[5]	M. R. Islam et al., "Smart parking management system to reduce congestion in urban area," in 2020 2nd international conference on electrical, control and instrumentation engineering (ICECIE), 2020: IEEE, pp. 1-6. 
[6]	N. T. Mai and D. M. Cuong, "The application of queueing theory in the parking lot: a literature review," in International conference on emerging challenges: Business transformation and circular economy (ICECH 2021), 2021: Atlantis Press, pp. 190-203. 
[7]	S. Madakam, R. Ramaswamy, and S. Tripathi, "Internet of Things (IoT): A literature review," Journal of computer and communications, vol. 3, no. 5, pp. 164-173, 2015. 
[8]	I. Ahmed, G. Jeon, and F. Piccialli, "From artificial intelligence to explainable artificial intelligence in industry 4.0: a survey on what, how, and where," IEEE transactions on industrial informatics, vol. 18, no. 8, pp. 5031-5042, 2022. 
[9]	J. Zhang and D. Tao, "Empowering things with intelligence: a survey of the progress, challenges, and opportunities in artificial intelligence of things," IEEE Internet of Things Journal, vol. 8, no. 10, pp. 7789-7817, 2020. 
[10]	V. K. Sarker, T. N. Gia, I. Ben Dhaou, and T. Westerlund, "Smart parking system with dynamic pricing, edge-cloud computing and lora," Sensors, vol. 20, no. 17, pp. 4669, 2020. 
[11]	A. Fahim, M. Hasan, and M. A. Chowdhury, "Smart parking systems: comprehensive review based on various aspects," Heliyon, vol. 7, no. 5, pp. 2021. 
[12]	M. Reda, A. Onsy, A. Y. Haikal, and A. Ghanbari, "Path planning algorithms in the autonomous driving system: A comprehensive review," Robotics and Autonomous Systems, vol. 174, pp. 104630, 2024. 
[13]	P. E. Hart, N. J. Nilsson, and B. Raphael, "A formal basis for the heuristic determination of minimum cost paths," IEEE transactions on Systems Science and Cybernetics, vol. 4, no. 2, pp. 100-107, 1968. 
[14]	M. Likhachev, G. J. Gordon, and S. Thrun, "ARA*: Anytime A* with provable bounds on sub-optimality," Advances in neural information processing systems, vol. 16, 2003. 
[15]	S. Choi and W. Yu, "Any-angle path planning on non-uniform costmaps," in 2011 IEEE International Conference on Robotics and Automation, 2011: IEEE, pp. 5615-5621. 
[16]	林晉嶽, "基於多攝影機交遞之智慧停車車輛追蹤系統," 淡江大學資訊工程學系碩士班學位論文, 2025. 
[17]	呂家成, "利用改良 A* 演算法之智慧停車場路徑規劃與效率優化," 淡江大學資訊工程學系碩士班學位論文, 2025. 
[18]	D. Berchmans and S. Kumar, "Optical character recognition: an overview and an insight," in 2014 International Conference on Control, Instrumentation, Communication and Computational Technologies (ICCICCT), 2014: IEEE, pp. 1361-1365. 
[19]	L. Jiao et al., "A survey of deep learning-based object detection," IEEE access, vol. 7, pp. 128837-128868, 2019. 
[20]	S. Y. Mohammed, "Architecture review: Two-stage and one-stage object detection," Franklin Open, pp. 100322, 2025. 
[21]	G. Jocher and J. Qiu, "Ultralytics YOLO11." https://github.com/ultralytics/ultralytics (Accessed August, 2025) 
[22]	M. L. Ali and Z. Zhang, "The YOLO framework: A comprehensive review of evolution, applications, and benchmarks in object detection," Computers, vol. 13, no. 12, pp. 336, 2024. 
[23]	J. Hosang, R. Benenson, and B. Schiele, "Learning non-maximum suppression," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 4507-4515. 
[24]	N. Bodla, B. Singh, R. Chellappa, and L. S. Davis, "Soft-NMS--improving object detection with one line of code," in Proceedings of the IEEE international conference on computer vision, 2017, pp. 5561-5569. 
[25]	N. Aharon, R. Orfaig, and B.-Z. Bobrovsky, "BoT-SORT: Robust associations multi-pedestrian tracking," arXiv preprint arXiv:2206.14651, 2022. 
[26]	Y. Zhang et al., "Bytetrack: Multi-object tracking by associating every detection box," in European conference on computer vision, 2022: Springer, pp. 1-21. 
[27]	PaddlePaddle, "PaddleOCR." https://github.com/PaddlePaddle/PaddleOCR (Accessed August, 2025) 
[28]	M. Liao, Z. Wan, C. Yao, K. Chen, and X. Bai, "Real-time scene text detection with differentiable binarization," in Proceedings of the AAAI conference on artificial intelligence, 2020, vol. 34, no. 07, pp. 11474-11481. 
[29]	B. Shi, X. Bai, and C. Yao, "An end-to-end trainable neural network for image-based sequence recognition and its application to scene text recognition," IEEE transactions on pattern analysis and machine intelligence, vol. 39, no. 11, pp. 2298-2304, 2016. 
[30]	A. Graves, S. Fernández, F. Gomez, and J. Schmidhuber, "Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural networks," in Proceedings of the 23rd international conference on Machine learning, 2006, pp. 369-376. 
[31]	R. Strukov and V. Athitsos, "Evaluation of video compression methods for network transmission on diverse data: A case study," in Proceedings of the 16th International Conference on Pervasive Technologies Related to Assistive Environments, 2023, pp. 300-305. 
[32]	S. Thrun and A. Bücken, "Integrating grid-based and topological maps for mobile robot navigation," in Proceedings of the national conference on artificial intelligence, 1996, pp. 944-951. 
[33]	E. W. Dijkstra, "A note on two problems in connexion with graphs," in Edsger Wybe Dijkstra: his life, work, and legacy, 2022, pp. 287-290. 
[34]	A. Buluç, S. Beamer, K. Madduri, K. Asanovic, and D. Patterson, "Distributed-memory breadth-first search on massive graphs," arXiv preprint arXiv:1705.04590, 2017. 
[35]	T. H. Cormen, C. E. Leiserson, R. L. Rivest, and C. Stein, Introduction to algorithms. MIT press, 2022. 
[36]	H. Zhu, S. Feng, and F. Yu, "Parking detection method based on finite-state machine and collaborative decision-making," IEEE Sensors Journal, vol. 18, no. 23, pp. 9829-9839, 2018. 
[37]	A. Kaur, S. Ayyagari, M. Mishra, and R. Thukral, "A literature review on device-to-device data exchange formats for iot applications," Journal of Intelligent Systems and Computing, vol. 1, no. 1, pp. 1-10, 2020. 
[38]	S. Dong, C. Cheng, and Y. Zhou, "Research on AJAX technology application in web development," in 2011 International Conference on E-Business and E-Government (ICEE), 2011: IEEE, pp. 1-3. 
[39]	M. Bostock, "ndjson-cli: Streaming newline-delimited JSON." https://github.com/mbostock/ndjson-cli (Accessed September, 2025)