視覺基礎之數化地籍圖影像處理

本文針對數化地籍圖影像，提出以計算視覺爲基礎之自動化處理程序，其中包括兩個主要演算法則：(1)國文分離，(2)網格至向量轉換。兩者均以簡單之串列爲資料結構，透過行列之一次掃瞄及線條追蹤程序，前者可獲得文字及符號的位置及大小，不僅完成圖文分離之處理，亦提供進一步文字辨識之資訊；後者可獲得地籍圖空間資料所需之位相關係：線條起點、中點、交點及終點等，配合線條的同步近似及同步邏輯補接，直線、折線及曲線之地籍圖元均可完整的予以重建和顯示。最後以六個常見之指標來檢定本文所提之演算法則，經實驗例測試，其效果頗佳。

關鍵字

地籍圖；圖文分離；行列掃描；網格至向量轉換；線條同步近似；同步邏輯補接

並列摘要

On the basis of computational vision, this paper presents an automatic process to focus on the image processing of digitalized cadastral map, The process includes two major algorithms: (1) segmentation (2) raster to vector conversion. Both of them use a simple list data structure to one-pass row majority scanning and line tracing. After the process, the former may obtain the posit ion, size of symbols and characters. Not only finishes the segmentation but also provides an information for the pattern recognition in the future; The latter may obtain the necessary topology to relate the spatial data of cadastral map, like line start point middle point intersect point and termainte point. To integrate with the synchronous approximation and the synchronous logical break-line linkage, straight line bend line and curve in cadastral map can be recon-structed and displayed completely. In the last, this paper ref res six index to confirm all alrorithms and concludes the effect very nice by the exprimental results.

並列關鍵字

Cadastral Map ； Segmentation ； Raster Scan ； Raster to Vector Conversion ； Line Synchronous Approximation ； Syncronous Logical Linkage

國際替代計量

視覺基礎之數化地籍圖影像處理

全文下載

主題瀏覽