目前擴增實境技術相關應用大部分以使用標記為主，但各式應用需求與日俱增，無標記(markerless)擴增實境技術使用上更具彈性，不必受限於標記的使用，因此應用層面更廣。視覺追蹤技術是擴增實境系統重要底層核心技術之一，但使用視覺追蹤技術在實際應用上易受到追蹤物件本身及外觀變化之影響，因此本文提出適用於無標記擴增實境應用之物件追蹤方法，能有效追蹤各式真實物件。首先框選設定追蹤物件；接著擷取物件特徵值，藉由特徵值比對以持續追蹤物件，並利用金字塔L-K光流法以縮短比對運算時間；最後經由2D-3D座標轉換，將3D虛擬物件疊加至真實場景。經由實驗證明本文所提之方法能有效追蹤各式真實物件，並能適用於無標記擴增實境相關應用。-Augmented Reality technology currently most relevant applications to use mark-based, but the increasing demand for all kinds of applications, without marking (markerless) amplification technology more flexible use of reality, not necessarily limited to the use of tags, therefore the application level broader. Visual tracking technology is one of the important augmented reality system underlying core technology, but the use of visual tracking technology in practical applications susceptible to track changes and appearance of the object itself, and therefore in this paper applies to Amplification unmarked objects of reality application tracing method can effectively track all kinds of real objects. First, set the track object marquee　then retrieve objects eigenvalues by eigenvalue than to keep track of objects, and uses of the pyramid LK optical flow method to shorten the computation time than on　and finally converted via 2D-3D coordinates of the 3D virtual object superimposed to the r