基于OpenCV的心电图像校准算法研究  

Research on ECG image calibration algorithm based on OpenCV

作　　者：曹鹏[1] 何雁萍 张志常[1] 常世杰[2] CAO Peng;HE Yanping;ZHANG Zhichang;CHANG Shijie(Department of Computer Science,School of Intelligent Medicine,China Medical University,Shenyang 110122,Liaoning Province,China;Department of Biomedical Engineering,China Medical University,Shenyang 110122,Liaoning Province,China)

机构地区：[1]中国医科大学智能医学学院计算机教研室,沈阳110122 [2]中国医科大学生物医学工程系,沈阳110122

出　　处：《中国数字医学》2023年第10期33-37,共5页China Digital Medicine

基　　金：辽宁省科学技术计划项目-基于内容的海量DICOM图像比对与检索方法研究(2021-YGJC-14)。

摘　　要：目的:基于OpenCV的心电图像校准算法,实现对纸质心电图拍摄图像的校准。方法:以不同背景、不同角度拍摄的纸质心电图像为研究对象,在Python和OpenCV环境下,采用图像预处理、边缘提取、轮廓检测等算法步骤,运用勾股定理和定点描绘分别进行图像校准。利用实拍图像进行对比实验,比较3种不同的边缘检测和两种不同的校准方法获取的图像效果,确定适用于纸质心电图拍摄图像的最优校准算法。结果:Canny算子在边缘检测中获取的图像效果最好,勾股定理方法可计算出原图像的边长并以较大精度校准还原图像,结果较为理想。结论:以Canny算子进行边缘检测和勾股定理计算边长的整体算法可用于纸质心电图拍摄图像的校准,这对医疗图像校准的后续相关拓展研究有一定的参考价值。Objective To realize the calibration of paper ECG(electrocardiogram)images based on OpenCV.Methods Paper ECG images taken from different backgrounds and angles were used as the research object.In Python and OpenCV environment,image preprocessing,edge extraction,contour detection,and other algorithms were used to calibrate the images by using Pythagorean theorem and fixed-point rendering.Through comparative experiments by using real-life images,three different edge detection methods and two different calibration methods were compared to determine the optimal calibration algorithm for paper ECG images.Results Canny operator can achieve the best image performance in the edge detection step.The Pythagorean theorem method can calculate the edge length of the original image and perform image calibration and restoration with greater accuracy,and the results were satisfactory.Conclusion The integrat algorithm for edge detection by Canny operator and edge length calculate by Pythagorean theorem can be used to calibrate the paper ECG images,which is valuable for further research on medical image calibration.

关 键 词：心电图 畸变图像 倾斜矫正 图片识别 

分 类 号：R540.41[医药卫生—心血管疾病] R319[医药卫生—内科学]