基於影像擷取之車位狀態偵測

機構典藏 > College of Engineering > Graduate Institute & Department of Computer Science and Information Engineering > Thesis > Item 987654321/87937

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/87937

Title:	基於影像擷取之車位狀態偵測
Other Titles:	Parking space status detection based on image capture
Authors:	盧孝宏;Lu, Hsiao-Hong
Contributors:	淡江大學資訊工程學系碩士班蔡憶佳
Keywords:	車位狀態偵測;色彩統計;角點偵測;Parking space detection;Statistics of color;Corner detection
Date:	2012
Issue Date:	2013-04-13 11:53:08 (UTC+8)
Abstract:	現今停車場的管理越來越受重視，能夠快速的導引駕駛到停車位上一直是最重要的課題，以至於停車格的狀態偵測是停車場監控管理系統最重視的技術之一，雖然已經有做法是直接在每個車位上安裝一個感測器來監控，但是這樣的方法不但耗費成本且維護費工。本篇論文提出利用停車場既有的監視器來擷取照片並且分析停車格的演算法，這樣的做法不僅可以大幅減少安裝的困難度，成本上來說也是非常節省的。本篇論文所提出的辦法是基於影像處理技術，主要利用色彩統計以及角點偵測來個別判斷停車格的狀態，再利用這兩種方法的數據資料來做為依據並且以投票方式來決定判斷的結果。我們在實際的停車場中模擬我們所提出的方法，模擬的場景包括含有車輛及未含有車輛時的各種情況，透過這樣的方式以貼近真實在停車場時可能會發生的各種情況，最後我們的測試結果顯示出本篇論文所提的車位狀態偵測方法準確率高達97%。 One important issue in parking lot management is to guide incoming cars to available parking space promptly. Therefore empty parking space detection is critical in producing efficient parking lot management system. This paper uses existing parking monitor to capture photos and analyzes the parking space, such an economic approach can significantly reduce the installation difficulty. The approach proposed in this thesis is based on image processing technology, color statistics, and corner detection to determine the parking space of the status of the individual, using the data of these two methods to do as the basis and a poll to determine the judgment results. We simulate our method in the parking lot. The simulation scenes contains variety situation of vehicles and without vehicles. Using this method can simulate several situations in a parking lot. The simulation results show the accuracy of our parking state detection method up to 97%.
Appears in Collections:	[Graduate Institute & Department of Computer Science and Information Engineering] Thesis

Files in This Item:

File	Size	Format
index.html	0Kb	HTML	175	View/Open

Loading...