摘要
针对磨粒显微图像识别中的目标提取问题 ,应用一种新的阈值分割方法将磨粒目标从复杂背景中分离。该算法以原始图像与分割后图像的相关系数最大为准则 ,选取最优阈值对磨粒图像进行二值化 ,经腐蚀、膨胀等滤波运算后 ,利用跟踪搜索法提取磨粒边界 ,实现了复杂背景下磨粒目标的有效提取。应用实例表明了算法的简洁有效性。
For object extraction inwear debris recognition system,a new thresholding method was applied to segment wear debris from its complex background. An optimum threshold was selected by maximizing the correlation between the original halftone image and the thresholded bi-levelimage. Having filtered by combining erosion and dilation operation,the contour of wear debris was obtained using boundary tracking method. Thus the extraction of wear debris objects wasrealized. Experimental results show that our method is feasible and effective.
出处
《机械科学与技术》
CSCD
北大核心
2004年第1期57-59,共3页
Mechanical Science and Technology for Aerospace Engineering
基金
国家自然科学基金项目 (5 0 175 0 69)资助
关键词
图像分割
阈值选取
磨粒
边界跟踪
Segmentation
Threshold selection
Wear debris
Boundary tracking
