主題樂園動態遊程規劃系統

Abstract

隨著主題樂園産業不斷發展，排隊問題成爲影響遊客遊玩體驗的主要因素，業界學界紛紛提出瞭解決方案，如Fastpass系統，旅遊行程規劃，但這些方案存在不足：1、過往研究模型不符合實際遊玩習慣。2、不能動態規劃線路。針對以上不足，本文提出以下三個研究目的1、建立合理的主題樂園靜態游程規劃模型。2、利用CCTV影像，自動獲取遊樂設施的實時排隊等待時間。3、發展出一個動態游程規劃模式，幫助游客實現實時動態游程規劃。 本文首先建立主題樂園靜態遊程數學規劃模型，並將模型求解分為兩個部分，分別為利用窮舉法求解必去點的最佳排序和利用逐個插空的方式選出最佳的候選點進行插入。然後，對CCTV影像進行逐幀提取、前景提取和灰度共生矩陣四個指標（能量、熵、均勻度、對比度）計算，之後與排隊人數進行多元線性回歸建模。最後，本文提出遊客狀態判斷、偏離道路檢測、等待時間變化檢測、統計模板匹配與準確度檢測四個部分組成的動態遊程規劃模型，并實現主題樂園動態遊程規劃App的建置，該系統包含基本的地圖功能、靜態遊程規劃功能、動態遊程規劃功能。 研究成果包括等待時間模板的選取與檢驗，影像判釋回歸模型建立與檢驗、動態遊程規劃App的建置、邏輯檢驗以及合理性檢驗。本文的動態遊程規劃模式、能夠幫助遊客實時動態更新行程時間和调整行程，減少遊客遊玩中總的遊玩時間，提高遊程規劃時間精度及遊客遊玩滿意度。

With the development of the theme park industry in recent year, queuing issues have become the main factors influencing tourists' play experience. There are some effective solutions such as the Fastpass system and tourist travel design problem. However, these solutions have deficiencies: 1. Past research models are not suitable for actual play. 2. there is not dynamical path planning function. In view of the above deficiencies, this paper proposes the following three research purposes: 1. Establish a rational theme park static tour planning model. 2. Using CCTV images to automatically obtain the actual waiting time for the amusement facilities. 3. Develop a dynamic tour planning model to help tourists achieve real-time dynamic travel planning. This paper firstly establishes a static tour planning model for the theme park, and divides the model solution into two parts, namely, using the exhaustive method to solve the best ordering of must-go points and selecting the best candidate points one by one to insert. Then, The frame-by-frame extraction, foreground extraction, and four-parameter calculation of Gray-Level Co-occurrence Matrix (energy, entropy, evenness, contrast) are performed on the CCTV image, and multiple linear regression modeling is performed with the queuing population. Finally, this paper proposes a dynamic tour planning model consisting of four parts: tourist state judgment, departure road detection, waiting time change detection, waiting time template matching and accuracy detection, and build the theme park dynamic tour planning App which include Basic map functions, static run planning functions, and dynamic run planning functions. The research results include the selection and testing of waiting time templates, the establishment and testing of image interpretation regression models, the construction of dynamic tour planning app, logical tests, and the rationality tests. The dynamic tour planning mode of this article can help tourists dynamically update the travel time and adjust the itinerary in real time, reduce the total playing time, improve the time precision of tour planning and the satisfaction.