索取号分割方法研究

摘要

拍摄得到的索取号照片通常含有很多干扰信息，需要进行相应的图像分割才能得到正确的索取号。在分割算法中，首先需取得原照片的灰度图像，然后经过多种算子对比后使用 CANNY算子来检测该灰度图的边缘点。检测得到的边缘点在经过S分量的进一步分割后，仍然还有少量的非文字边缘点干扰。为了排除这些干扰，采用文字行、列区域检测来确定边缘点数量最多的区域，从而正确的分割出照片中的索取号部分。%Call numbers’ photos obtained by shooting usually had interference information. So image segmentation was an indispensable means of obtaining exact call numbers. Firstly, call numbers’ photos should be converted to gray level images. Secondly, using CANNY operator detected this images edge points after comparing with several kinds of operators. But a little interference information was still existed among the words’ edge points after using S component segmentation. Finally, we used edge area detection which based on row and column to eliminate the interference and segment exact call numbers’ edge.