以控制直線進行影像外方位參數求解之自動化系統

Abstract

傳統攝影測量空中三角作業主要以量測點特徵(Point Features)為主，此類特徵必須在影像中呈現足夠的訊號，相對於此，線型特徵(Linear Features)具有較清楚的幾何及輻射訊息，也因此提供偵測及匹配上的便利。 本研究利用從舊有數值圖資中萃取出之直線特徵作為控制資料，嘗試利用自動匹配之方式找尋物、像空間直線特徵的對應關係，並以直線為基礎之共線式(Line-based Collinearity Equations)求解像片外方位參數。在實際資料的測試中，以平差計算成果、共軛光束一致性及外部精度的檢查，証實了以控制直線求解像片外方位參數的可行性及合理性。

Point features with distinctive radiometric signal have been commonly accepted as major measurements for photogrammetric aerial-triangulation. Linear features, on the other hand, rendering rich geometric as well as radiometric information of the scene, would offer great advantage for automatic detection and matching tasks. This study employs control lines derived from existing geo-spatial data for solving photo orientation. An automatic line feature matching scheme between object space and image space has been developed, followed by line-based collineaity equations for orientation parameter estimation taking error detection into account. Both the simulated and experimental test results reveal that the proposed approaches are applicable and the achievable accuracy level is satisfactory for the task of photo orientation.