【摘要】 手绘草图图像的前期分割处理对提高手绘草图识别质量具有重要作用.提出了一种复杂草图符号的分割算法,该方法采用霍夫直线检测去除干扰的导线,再进行屏幕搜索.利用连通域等特性,分割出各个待识别的电气元器件符号.改进的连通域标记算法可在一次扫描图像的条件下,完成对像素的标记,可缩短算法的执行时间.对于连通域数目较多的图像,时间缩减效果更明显.实验结果表明,该方法能够自动分割出合适图元单位的区域,实用性强.更多还原

【Abstract】 The initial segmentation of the hand-drawn sketch image plays an important role in improving the quality of the hand-drawn sketch recognition.An algorithm for the segmentation of complex sketch symbols has been proposed.The method uses Hough line detection algorithm to remove the interference of the wire,and carries on the screen search.By using the characteristics of the connected component,the segmentation of the various electrical components have been identified.The improved connected component labeling algorithm can mark pixels in the condition of scanned image once,which can shorten the execution time of the algorithm well,and the time reduction effect is more obvious for the image with a large number of connected domains.Experimental results show that the proposed method can automatically segment the region of the appropriate unit,and the algorithm is effective and practicable.更多还原