机构地区: 新疆师范大学物理与电子工程学院
出 处: 《新疆师范大学学报(自然科学版)》 2012年第4期42-45,共4页
摘 要: 自然环境中包含大量的文字信息,这些信息对于基于内容的图像检索、场景理解和智能交通等领域具有重要的应用价值。文章在现有文字定位技术和分析维吾尔文字母结构特点的基础上,提出了用Sobel边缘检测的一种定位方法,并详细介绍了用边缘检测、二值化、形态处理、最后利用连通区域分析和文本区域合并等算法来提取文本区域的过程。 The natural environment contains plenty of text information which has an important applica- ble value in the fields-image retrieval based on content, scenes understanding and intelligent transportation etc. Based on existing text location technology and analyzing the structure features of Uyghur character, the thesis proposes a kind of text location method using Sobel edge detect. The thesis will also introduce the process of extracting text area in detail, by firstly using algorithm of edge detection.binaryzation. morphological processing, and finally using the algorithm of connected component analysis and text region merging etc.