基于梯度霍夫变换的藏文手写采样表格检测

Table Detection of Tibetan Handwritten Sample Forms Based on Gradient Hough Transform

藏文手写字丁的提取是藏文手写字丁特征分析的基础,是脱机藏文手写体识别预处理工作的重要部分.为了能够顺利提取藏文手写采样表格中的藏文字丁,我们对现有的表格检测方法加以分析.在本文中,利用霍夫变换检测直线的原理,我们对基于梯度的霍夫变换算法进行改进,将其应用到藏文手写体采样表格中表格线的检测.试验证明,该算法能够较好地检测出采样表格中的直线.

Extracted tibetan handwritten character is basis of Tibetan handwritten character analysis.It is also an importont part of off-line Tibetan handwritten character preprocessing.In order to successfully extract Tibetan handwritten character from sample form,the existing table detection methods were analyzed.The principle of Hough transform to detect lines is utilized.Gradient based Hough transform algorithm was improved and is applied to the table detection of Tibetan handwritten sample forms.The performance of the algorithm is demonstrated by experiments.