气泡高干扰在线铁谱视频图像的磨粒快速分割算法研究

针对在线铁谱视频图像气泡高干扰所面临的磨粒分割困难问题，提出一种气泡高干扰在线铁谱视频图像的磨粒快速分割算法。首先运用运动检测的方法确定视频中气泡的位置，并用相邻帧相同位置的图像信息对气泡区域进行处理，再使用双边滤波对处理后的图像进行平滑去噪，实现气泡干扰的初步抑制；最后基于抑制气泡图像的灰度直方图，对每一帧图像选取其自适应的阈值，实现在线铁谱视频图像中磨粒的快速分割。该研究为在线铁谱的磨粒分割与后续对磨粒特征的智能提取和分析奠定了基础。

Study on Fast Wears Segmentation Algorithm for Online Ferrography Video Images in High Interference of Bubble

A fast wear particle segmentation algorithm based on high bubble interference was proposed to solve the difficult problem of wear particle segmentation in online ferrography video images.Firstly,the location of bubbles in the video was determined by motion detection method,and the area of bubbles was processed by the image information in the same position of adjacent frames.Then,the processed image was smoothed and denoised by bilateral filtering to realize the initial suppression of bubble interference.Finally,based on the gray histogram of the images after bubble elimination,the adaptive threshold of each frame was selected to realize the fast segmentation of wear particles in on line ferrography video images.This study lays a foundation for on line ferrography wear particle segmentation and the next intelligent extraction or analysis of wear particles characteristics.