基于多复杂关键词的模板图像匹配矫正算法  

Template Image Matching and Correction Algorithm Based on Multiple Complex Keywords

作　　者：张昊[1] 刘义亭 李佩娟[1] ZHANG Hao;LIU Yiting;LI Peijuan(Nanjing Institute of Technology,Nanjing,Jiangsu 211167,China)

机构地区：[1]南京工程学院,江苏南京211167

出　　处：《自动化应用》2024年第20期165-167,170,共4页Automation Application

摘　　要：针对当前智能电表示数在文字识别时图像矫正效果不佳导致识别模板无法生成、识别结果精度较低等问题,提出了一种基于多复杂关键词的模板图像匹配矫正算法。该算法通过选择合适的关键词,利用标注软件提取关键词的坐标实现智能电表的模板图像匹配矫正,为电表后续的示数识别提供了良好的识别模板。通过后续多种智能电表识别实验,证明该算法具有简单方便、鲁棒性强、矫正效果好和矫正成功率高等特点。与基于四点定位的图像矫正算法相比,所提算法矫正成功率的提升超过20%。A template image matching correction algorithm based on multiple complex keywords is proposed to address the problems of poor image correction effect of smart meter readings in text recognition,resulting in the inability to generate recognition templates and low accuracy of recognition results.This algorithm selects appropriate keywords and uses annotation software to extract the coordinates of the keywords to achieve template image matching correction for smart meters,providing a good recognition template for subsequent meter reading recognition.Through various subsequent smart meter recognition experiments,it has been proven that this algorithm has the characteristics of simplicity,convenience,strong robustness,good correction effect,and high correction success rate.Compared with the image correction algorithm based on four point positioning,the proposed algorithm improves the correction success rate by more than 20%.

关 键 词：模板匹配 OCR识别 初始模板 

分 类 号：U495[交通运输工程—交通运输规划与管理]