针对地形扫描图中高程数据的特点,提出了一种高程数据的自动探测与识别方法。首先,根据数学形态学中提取连通分量的算法探测并分离高程数据;其次,采用环投影变换及分形维数得到高程数据中数字字符的旋转不变性特征向量,并利用该特征向量识别高程数据。实验表明,该方法能准确地识别高程数据。
According to the features of the elevation values in scanned topographical maps, an approach to detect and recognize elevation values was presented. Firstly, an algorithm for extraction of connected components based on mathematical morphology was employed to detect and separation elevation values. Secondly, by use of ring-projection transformation and fractal dimension, the characteristic vectors of the digital characters in elevation values that were invariant to rotation was obtained. Finally, the elevation values were recognized by use of the characteristic vectors. Experimental results show that the approach can accurately identify the elevation values.