摘要
针对票据图像的特征, 本文提出了一种票据字符识别的预处理算法。采用基于方向投影的倾斜校正方法来调整图像的倾斜, 再经过最佳阈值分割得到二值化图像, 去除离散杂点后, 采用改进后的基于方向投影的方法进行字符分割。实验表明, 该算法能够满足快速准确的字符分割, 具有很强的抗干扰性和应用适用性。
On the base of analyzing printed bill image, in this paper, a new preprocessing algorithm is proposed to meet needs of high real-time of the high-speed image processing in the on-line recognize of printed symbols, which is a improved skew correction based on directional projection , an optimal threshold division algorithm is adopted to get binary image and a contour projection based character segmentation algorithm is used to get single number image. Experiments show that our methods can segment each single printed number of an image rapidly and accurately, and also have high noise endurance and application adaptability.
出处
《微计算机信息》
北大核心
2007年第18期261-263,共3页
Control & Automation
基金
国家自然科学基金(资助号:60474021)
湖北省科技攻关计划(资助号:2004AA101B01)
武汉市青年科技晨光计划(资助号:20045006071-26)的资助
关键词
字符识别
预处理算法
倾斜校正
阈值分割
方向投影
Character Recognition, Preprocessing Algorithm, Skew Correction
Threshold segmentation, Directional Projection
