机器视觉在小型管道探伤中的应用研究

发布时间：2018-03-18 03:37

  本文选题：管道探伤　切入点：管道机器人　出处：《化工自动化及仪表》2016年11期 　论文类型：期刊论文

【摘要】：设计了针对小型管道内部缺陷检测的螺旋管道机器人系统,基于该机器人系统提出了图像处理的改进算法。首先采用结合中值滤波思想的双边滤波器,解决了双边滤波无法去除孤立噪声点的问题;其次采用了二维最大熵的阈值分割方法进行图像分割;最后根据管道缺陷的特点提取适合分类器分类的代表特征点进行分类。仿真研究表明:所提算法能够更加完整地提取缺陷信息。
[Abstract]:A spiral pipeline robot system is designed to detect the internal defects of small pipelines. Based on the robot system, an improved image processing algorithm is proposed. Firstly, a two-sided filter based on median filter is used. The problem that bilateral filtering can not remove isolated noise points is solved. Secondly, the threshold segmentation method based on 2-D maximum entropy is used for image segmentation. Finally, the representative feature points suitable for classifier classification are extracted according to the characteristics of pipeline defects. The simulation results show that the proposed algorithm can extract defect information more completely.
【作者单位】： 天津理工大学天津市复杂系统控制理论及应用重点实验室;
【基金】：天津市科技重大专项与工程项目(15ZXZNGX00140) 天津市应用基础与前沿技术研究计划项目(16JCTPJC49400)
【分类号】：TP391.41
，

本文编号：1627864