摘要： 视频图像中包含着许多重要的文字信息。图像和视频文本信息的提取包括文本检测、定位、跟踪、提取、增强和识别等几个部分。将文本的检测、定位与提取，作为文本区域提取的整体来讨论。以文本的检测算法为重点研究对象，提出了应用小波模极大值算法来解决视频图像中文本区域的检测。实验表明，小波模极大值算法所得到的文本区域与其它算法相比具有更好的评价指标。

关键词: 文本提取, 小波模极大值, 滑动窗口

Abstract: The text in picture and video includes many useful information. Extraction of text information in picture and video includes text detection, location, tracing extraction, image enhancement and character recognition. Regarding the detection, location and extraction as a whole, the algorithm of the text detection is the important target of the research, it puts forward of the wavelet modulus maximum (WMM) as the function to solve the detection of the text in the videos. The experiment expresses that the result of the WMM is better than other algorithms.

Key words: Text extraction, Wavelet modulus maximum(WMM), Slip window