摘要
介绍了一种基于面向对象的Visual C++语言,在Windows MFC平台下开发出的视频字幕自动提取系统。该系统应用垂直、水平、对角方向的边缘检测算子检测出3个方向的字幕边缘信息,然后运用形态学对每个方向的边缘图像进行处理,最终运用与融合提取出字幕区域。实验结果验证,该系统性能稳定,字幕定位较精准。
This paper introduces a system for automatic text extraction in video images, which is based on Windows MFC with object-oriented language---Visual C＋＋. First, this system uses three directions of edge detection operator to detect text edge information in three directions, then uses morphology for each direction of edge images. Finally, it uses “and” mergeing for text regions extraction. Experiments show that this system achieves good effect in property and text extraction.
出处
《微型机与应用》
2014年第20期38-40,45,共4页
Microcomputer & Its Applications
