VISION-BASED IMAGE-PROCESSING OF DIGITIZED CADASTRAL MAPS

This paper presents an automatic method for processing digitized image s of cadastral maps. The method includes two major algorithms: a segme ntation and a Raster-to-Vector conversion. Those algorithms use a simp le data-list structure for recording data created during single-pass, row-majority scanning and line tracing. The segmentation algorithm obt ains the positions and sizes of symbols and characters, in addition to completing map segmentation and proving useful for pattern recognitio n. The Raster-to-Vector conversion algorithm obtains topological infor mation necessary to relate cadastral map spatial data to line start po ints, midpoints, intersection points, and termination points. It consi sts of four integrated sub-algorithms that remove noise, unify run-len gth coordinates, and perform synchronous line approximations and logic al linkage of line breaks. Straight, angled, and curved lines can then be completely reconstructed for display. Also presented are six indic es that verify algorithm and experimental results.