一种多传感器云融合技术的亚面表缺陷深度检测算法研究

摘要

针对未知深度的亚表面缺陷检测,各种传感器产生的特征信息难以直接应用的难题,提出了一种多传感器云融合技术的亚表面缺陷深度检测算法.首先,本文简要介绍涡流检测与超声检测两种经典无损检测模式;然后,在已知缺陷深度的样本集上建立云滴数学模型获取单一测量模式下缺陷深度的隶属度分布函数;最后,提出了一种D-S+PCR信息融合算法将不同传感器获取的隶属度分布转化为概率分布进行数据融合,以高斯拟合的方式实现亚表面缺陷深度的检测估计.实验结果显示:对比传统Bayes变换的D-S证据融合检测技术与单传感检测技术,本文提出的检测算法对亚表面缺陷深度检测具有较高的识别准确率.%In consideration of the difficulty of directly using the multi-sensor detecting features information for the defect identification. The paper proposes an improved multi-sensors recognition algorithm based on cloud-fusion technology for subsurface defect depth evaluation. At first,two common nondestructive testing technologies such as ultrasonic testing( UT) ,eddy current testing( ECT) are introduced;at second the cloud model to calculate the proba-bility distribution of single detecting method is carried out;then a fusion algorithm based on D-S and PCR theory is improved and used to fuse the probability from transformation the membership of multi-sensors,at last Gaussian fit-ting method is taken to evaluate the subsurface defects depth. The experimental result shows that the improved algo-rithm is superior to the existing algorithm;it can achieve better synthesis results and improve the correct target rec-ognition rate.