According to the morphic characteristics of the Chinese characters and background areas in a complex background image, we propose a fast text segmentation algorithm based on morph operation, and compare it with the LLT algorithm.

• 本文针对复杂背景图像中文字与背景区域的形态特征,提出了一种基于形态运算和区域生长的快速文字分割算法MOSA,同时将之与LLT分割算法进行了比较。