机器视觉在小型管道探伤中的应用研究

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。

Application Research of Machine Vision in Flaw Detection of Small-size Pipes

A robot system for spiral pipes was designed to detect internal defects in small -size pipes , and the improved algorithm based on the robot system was proposed for image processing .Firstly, having the bilateral filter which combined with median filtering adopted to remove isolated noise points; secondly , having the threshold segmentation method of 2D maximum entropy taken for image segmentation and finally ,having the characteristics of pipe defects based to extract their feature points for classification .Simulation results show that the proposed algorithm can extract defects ’ information more completely .