扫描地形图中等高线的识别(英文)

Contour lines recognition from scanned topographic maps

针对一般质量地形图的扫描过程中颜色混淆和错误现象比较严重,以致于在彩色分割等高线时不仅由于其自身与其他地图符号交叉或交叠而形成的断裂变得更为复杂,而且等高线间的密集排列也会造成粗线现象(等高线彼此粘连在一起)等问题提出一种有效的针对一般质量地形图中等高线的识别方法。该方法具有3个特点:一是彩色地图分割中使用关键颜色集技术以补偿混淆和错误颜色;二是在原图像中等高线出现断裂和粗线的地方采用局部窗口再分割算法,以解决等高线断裂和彼此粘连问题;最后利用等高线的全局拓扑性质滤波得到正确的等高线。实验结果表明该方法对于一些高质量和一般质量的地形图中等高线的识别都具有一定的应用潜力。

Aliasing and false colors in the scan of a common-conditioned map are so serious that when contour lines is segmented from it, not only many gaps due to contour lines intersecting and/or overlapping with other features become more intricate, but also thick lines （lines conglutination） due to contour lines closely spacing occur. An effective method is presented for contour lines recognition from scanned common-conditioned topographic maps. It consists of three distinguished aspects. Firstly, a color key set technique is adopted in color map segmentation to compensate for aliasing and false colors. Secondly, local window segmentations are performed at the place where the gaps and thick lines occur in the original image to resolve contour lines disconnection and conglutination. Finally, the global topologic properties of contour lines are used to filter the accurate contour lines. The experimental results show that this new method is potential when applied to a number of well-conditioned and common-conditioned topographic maps.