對於含單一攝影機及數個標誌點的視覺系統，提出與其對等類比的空間機構，並進而加以解析。物體與相機如同空間機構的移動平台與基座，標誌點則相似於可伸縮的支撐腿。藉此對等機構，除可對視覺系統提供另一種解讀方式，亦可使兩個主要問題迎刃而解。其一為若欲成功重建物體的姿態，所需標誌點的最低數目；該問題轉化為空間機構的自由度分析。另一為從單一張影像的資訊以重建物體的姿態；此則如同於對機構作構形分析。結果顯示俱六個自由度的物體，最少需四個標誌點始能重建其姿態；且若其中三點共線，計算過程可大幅簡化。對於使用不同數目的標誌點以搭配物體所受的各類拘束，亦做分析及探討。相關模式、概念與成果，亦可擴及至由多個物體相連結的系統。
A vision system that includes a single camera and several landmarks attached to the objects is studied using an analogous spatial mechanism. The object is equivalent to the moving platform, while the camera and landmarks are modeled into the fixed platform and supporting legs respectively. The proposed analogous model gives landmark-based vision system an alternative interpretation, and makes it easy to solve two primary problems. First is the determination of the minimum number of landmarks so that the pose of an object can be reconstructed correctly. This number is obtained by calculating the degrees of freedom of the mechanism. The second problem is reconstructing the object’s pose analytically from a single image, and similar to analyzing the analogous mechanism’s configuration. Four landmarks are shown necessary for reconstructing the pose of an object with six degrees of freedom, and the computation becomes straightforward while three of them are collinear. The strategies of attaching landmarks properly and reconstructing pose efficiently for objects being subject to various movement constraints are studied. The results also extend to the application for multi-body systems.
