本篇論文描述一個方便的影像繪圖的方式,由多張二維參考輪廓影像去計算在某個視角下,螢幕所要描繪三維影像空間資訊的方法,本研究可以應用在數位典藏方面。 以多張二維參考輪廓影像計算彼此影像之間位置的關係,以視覺外型(visual hull)技術計算三維影像空間資訊。此方法利用極線幾何(epipolar geometry)、Plücker coordinate,藉由世界座標系統、視窗座標系統與OpenGL螢幕座標系統之間的轉換求出兩張圖片的幾何關係與對應點,最後依上述等方法為基礎,進行視覺外型的計算。 結果的精細程度和參考影像的數量有關,由兩張參考影像所求的結果距離真正物體的外型還有一段差距,而六張參考影像所得到的結果,就越接近真正物體的結果。
In this paper, it describes an efficient image-based approach to computing the scene geometry described by visual hulls – shapes obtained from silhouette image data. A visual hull is constructed by using the visible silhouette information from multiple reference images. If implemented successfully, it could be useful in digital archive. We use multiple reference images to compute the scene geometry by using visual hull approach. The algorithm takes advantage of epipolar geometry and Plücker coordinate system to compute the geometry relationship and feature point of two images with world coordinate system, screen coordinate system and view coordinate system. The experiment all results show that the apparent quality is not close to the object’s shape when it is shaded with the view-dependent texture from two reference images; nevertheless, the six reference images of an object give a strong indication of that object’s shape.