博碩士論文 87325061 完整後設資料紀錄

DC 欄位	值	語言
DC.contributor	資訊工程學系	zh_TW
DC.creator	徐凱輬	zh_TW
DC.creator	Kai-liang Hsu	en_US
dc.date.accessioned	2000-7-10T07:39:07Z
dc.date.available	2000-7-10T07:39:07Z
dc.date.issued	2000
dc.identifier.uri	http://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=87325061
dc.contributor.department	資訊工程學系	zh_TW
DC.description	國立中央大學	zh_TW
DC.description	National Central University	en_US
dc.description.abstract	經由一連串連續動作影像的分析來判定物體移動的速度，再藉此速度來求得物體與偵測器間的距離是本研究的目標。為了達此目標，在本論文的研究中，我們發展一個結合了光流偵測 (optical flow estimation) 與深度計算 (range determination) 的被動式距離檢知即時系統。在光流偵測中，我們採用兩個不同性質的方法來估計影像中的光流值；一是Horn-Schunck方法；二是 Lucas-Kanade方法。在Horn-Schunck方法中，我們加入多重解析度的處理以加快光流計算的收斂速度，並得以使本系統更加符合即時的要求。藉由已知的光流資訊，配合有少許動作限制的透視投影公式，我們就能夠計算出目標物與偵測器間的距離，而此項距離資訊將可應用於自動飛行導航系統和軍事戰術演習上。	zh_TW
dc.description.abstract	In the applications of autonomous flight navigation and tactical simulation, we need the range information acquired from the motion cues. In this study, we develop optical flow estimation methods as well as a range determination method to achieve the purpose. Two methods for estimating optical flow from image sequences are compared; one is the Horn-Schunck method and the other is the Lucas-Kanade method. Based on the motion data of optical flow and the principle of perspective transformation, the distance from the targets to the sensor is determined. In the optical flow estimation, the multiresolution strategy is utilized to speed up the process to make our system being suitable for real-time applications.	en_US
DC.subject	深度計算	zh_TW
DC.subject	多重解析度	zh_TW
DC.subject	光流	zh_TW
DC.subject	Range Determination	en_US
DC.subject	Multiresolution	en_US
DC.subject	Optical Flow	en_US
DC.title	多重解析度光流分析與深度計算	zh_TW
dc.language.iso	zh-TW	zh-TW
DC.title	Multiresolution Optical Flow Estimation and Range Determination	en_US
DC.type	博碩士論文	zh_TW
DC.type	thesis	en_US
DC.publisher	National Central University	en_US