复杂背景图像中维吾尔文字切分与识别技术的研究  

Study on Character Segmentation and Recognition Technology of Uyghur in Image with Complex Background

作　　者：阿地力.依米提 刘吉超[1] 杜力坤.苏来曼 

机构地区：[1]新疆师范大学物理与电子工程学院,新疆乌鲁木齐830054

出　　处：《新疆师范大学学报（自然科学版）》2014年第1期65-68,2,共4页Journal of Xinjiang Normal University(Natural Sciences Edition)

基　　金：新疆少数民族科技人才特殊培养计划科研资助项目(201123114)

摘　　要：文字识别是图像处理和模式识别的主要研究方向之一。文章提出了一种复杂背景图像中维吾尔字符切分与识别方法,该方法根据维吾尔文字母结构特点,对维吾尔文词语的切分进行了认真地分析,通过对维吾尔文字符边缘特征和笔画分布规律等特点进行研究,提出了比较有效的切分方法,并结合维吾尔文的特有规则,利用维吾尔文周边结构模式识别方法提取结构特征,最后通过计算Minkowski距离找出最近的类别,输出识别结果。Character recognition is one of the research directions of the image processing and pattern recognition .This paper presents a Uyghur character segmentation and recognition method in image with complex background .According to the structural features of Uyghur alphabet ,we analyzed the Uyghur text segmentation carefully .Through studied on the edge characteristics and strokes distribution of Uy-ghur ,the better segmentation method has been put forward .Combined with the specific rule of Uyghur , the structural characteristics were extracted by the surrounding structure recognition methods of Uighur . Minkowski distance is calculated to find the nearest category and the recognition result is output at last .

关 键 词：切分 识别 维吾尔文 结构特征 

分 类 号：TN911.7[电子电信—通信与信息系统]