基于差商的油液监测铁谱图像自适应分割

发布时间：2019-05-25 03:16

【摘要】：针对油液监测中铁谱磨粒图像分割阈值难以选取的问题,本文提出一种基于差商的自适应铁谱图像分割算法。首先,将铁谱磨粒灰度图像转换成三维灰度直方图,并对其进行切片分析;然后,引入Newton插值多项式,将不同切片所得的频数作为切片灰度-频数曲线的插值点,基于差商构造第一类可接受函数和第二类可接受函数,结合实验数据确定两类误差,选取同时满足两类误差的最小灰度值作为分割阈值;最后,用本文方法对不同类型的磨粒图像以及添加高斯噪声和椒盐噪声后图像分别进行分割实验,并与经典的迭代阈值法、Otsu算法、最大熵法进行了比较。实验结果表明,本文方法受噪声干扰较小,误检率和漏检率整体优于其他3种算法。对分割所得的磨粒图像进行特征提取,并利用支持向量机进行识别,本文方法对3种故障磨粒识别准确率最高,达到82.86%,虽在运行时间上无明显优势,但综合性能最优,能满足油液监测过程中铁谱图像自适应分割的需求。
[Abstract]:In order to solve the problem that it is difficult to select the threshold of ferrography abrasive image segmentation in oil monitoring, an adaptive ferrography image segmentation algorithm based on difference quotient is proposed in this paper. Firstly, the gray image of ferrography abrasive particles is converted into three-dimensional gray histogram, and the slice analysis is carried out. Then, the Newton interpolation polynomial is introduced, and the frequency obtained from different slices is taken as the interpolation point of the gray-frequency curve of the slice. The first kind of acceptable function and the second kind of acceptable function are constructed based on the difference quotient, and the two kinds of errors are determined by the experimental data. The minimum gray value which satisfies the two kinds of errors at the same time is selected as the segmentation threshold. Finally, the segmentation experiments of different types of abrasive images and images with Gao Si noise and pepper and salt noise are carried out by using this method, and compared with the classical iterative threshold method, Otsu algorithm and maximum entropy method. The experimental results show that the proposed method is less interfered by noise, and the false detection rate and missed detection rate are better than the other three algorithms as a whole. The feature extraction of the segmented abrasive image is carried out, and the support vector machine is used to identify the abrasive particles. The accuracy of this method is the highest, up to 82.86%. Although there is no obvious advantage in the running time, the comprehensive performance is the best. It can meet the needs of adaptive segmentation of ferrography image in the process of oil monitoring.
【作者单位】：西安交通大学智能仪器与监测诊断研究所;新疆大学机械工程学院;
【基金】：国家自然科学基金资助项目(No.51421004,No.51365051) 教育部新世纪优秀人才支持项目(No.NCET-13-0461)
【分类号】：TP391.41

【相似文献】