基于HALCON平台快速开发光学字符识别系统的研究

HALCON平台提供了成熟的算法包、开放式结构的功能模块及光学字符识别助手,有利于快速开发光学字符识别系统,具有功能强大、开发成本低及操作简单的优点。文章应用了平台自带的多个分类器模板对字母和数字进行识别测试,结果证明HALCON平台自带分类器模板对部分大写字母和数字的识别具有较高的准确率。在平台上创建了开放式的分类器模板,建立样本训练集文件,对分类器进行训练,使用训练后的分类器模板对汉字进行识别测试,识别准确率也较高。

The HALCON Platform provides mature Algorithm package,open structure function module and optical character recognition assistant,which is helpful to develop the optical character recognition system,and has the advantages of powerful function,low cost and simple operation.The test results show that the classifier template of Halcon Platform has a high accuracy rate for recognizing part of the majuscule and numbers.An Open classifier template is created on the platform,and a sample training set file is set up to train the classifier.After training,the classifier template is used for Chinese character recognition test,and the recognition accuracy is also high.