這幾年3D技術已經慢慢普及化,不管是電影、遊戲、列印、消費性的電子產品等,都可以見到3D技術的應用。微軟公司開發的Kinect,取代手持或是踩踏控制器,使得3D技術的應用越來越廣泛,而三維重建主要在考古學、古物保存、醫學影像等領域被廣泛的應用。Kinect是配合紅外線發射器及紅外線CMOS攝影機直接擷取到深度的影像資訊,價格便宜許多且獲得的資訊更簡便更快速。本論文使用兩台Kinect當做擷取的工具,兩幅影像及多幅影像,可以直接擷取到彩色影像及深度資訊,再利用Harris角點偵測搜尋兩張圖的特徵點進行影像的匹配。點雲是由點所構成,每一個點都包含多種資訊,如RGB、深度、角點等,運用這些點的資訊可以在三維空間中重建物體及其環境。
We have seen many applications of 3D technology using in games, movies, and consumer electronic products.Microsoft released the Kinect which lower the threshold of 3D technology. The Kinect can operate easily without hand-held controller or a stampede. So the application of 3D technology more widely used, such as three-dimensional reconstruction in medical imaging, archeology, the preservation of antiquities.The principle of Kinect is used by an infrared emitter with CMOS camera and infrared images directly to a depth of information retrieved. Kinect is much cheaper, easier and more quickly to obtain information.In this thesis, we use two Kinects as a tool to take two images or multiple images that can capture the information directly to color images and depth images. Then, we use the method of Harris corner detection to find the feature points between two images. Point cloud are composed by many points, each point contains a variety of information, such as RGB, depth, corners, etc., information on those points can reconstruct the object and its environment in three dimensions.