节点文献

轨道缺陷图像检测算法的优化

Optimization of image detection algorithm for track defect

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 闵锐黄楚云张晓峻孙晶华

【Author】 MIN Rui;HUANG Chu-yun;ZHANG Xiao-jun;SUN Jing-hua;School of Science,Hubei University of Technology;College of Physics and Optoelectronic Engineering,Harbin Engineering University;

【通讯作者】 孙晶华;

【机构】 湖北工业大学理学院哈尔滨工程大学物理与光电工程学院

【摘要】 根据轨道健康状况无损检测的需求,基于OpenCV软件对轨道缺陷图像检测BM3D算法进行了优化.利用小波变换优化BM3D算法,提出BM2D+1D的算法模式,改进后算法的运算效率提高了39.6%,图像峰值信噪比提高了0.388dB.利用图像增强LOG算子和Gamma算子对Canny算法进行改进,不仅优化了算法运算复杂度,同时滤除了大部分虚假边缘信息和无意义像素信息,运算效率提高了81.4%,使Canny算法的阈值可大范围调节并确定最优解,提高了算法的灵活度.

【Abstract】 According to the requirements of track health non-destructive testing,the track defect detection BM3Dalgorithm was optimized based on OpenCV software.Wavelet transform was used to optimize BM3Dalgorithm,and the algorithm mode of BM2D+1Dwas proposed.The operation efficiency was improved by 39.6%,and the image peak signal-to-noise ratio was improved by 0.388dB.Canny algorithm was improved by using image enhancement operators,LOG operator and Gamma operator,which not only optimized the time complexity of the algorithm,but also filtered out most of the false edge information and meaningless pixel information.The operation efficiency was improved by 81.4%.The threshold of Canny algorithm could be adjusted in a large range to determine the optimal solution,which improved the flexibility of the algorithm.

【基金】 国家自然科学基金面上项目(No.61377084)
  • 【文献出处】 物理实验 ,Physics Experimentation , 编辑部邮箱 ,2022年02期
  • 【分类号】TP391.41;U216.3
  • 【下载频次】97
节点文献中: 

本文链接的文献网络图示:

本文的引文网络