一种改进的小波变换域的字符定位方法

An improved algorithm of character localization based on wavelet transform domain

在分析图像中字符特点的基础上,研究其在小波变换域中的特征,依据这些特征,分析频率加权阈值法和利用块间信息的平均阈值法的优缺点,将二者结合进行图像字符定位。实验表明,该算法可进行图像中的字符定位,同时又可以弥补两种方法定位时的不足。

On the basis of analyzing characteristics in images, the paper studies its wavelet transform domain. According to these characteristics, the paper analysis the advantages and disadvantages of frequency weighted threshold value method and utilization of information between blocks average threshold value method, and combines them to do the image character positioning. Experiments show that this algorithm can undertake image of character positioning; At the same time, also can make up for the two methods lack when positioning.