可视化与非可视化特征融合超声3D目标识别研究  

Research on Ultrasonic 3D Target Recognition Based on Visual and Non-Visual Feature Fusion

作　　者：宋寿鹏[1] 申静静 卢翠娥[1] SONG Shoupeng;SHEN Jingjing;LU Cuie(School of Mechanical Engineering,Jiangsu University,Zhenjiang 212013,China)

机构地区：[1]江苏大学机械工程学院,江苏镇江212013

出　　处：《电子科技》2019年第5期5-10,共6页Electronic Science and Technology

基　　金：国家自然科学基金(51375217)~~

摘　　要：目前常用的超声3D目标识别方法主要是利用传感器在空间一点或多点获取一维回波,通过信号处理得到目标体3D信息以实现3D目标体识别。这些方法普遍存在识别率低和鲁棒性差的问题,制约了该项技术的推广和应用。为此,文中提出了一种基于可视化和非可视化特征融合的超声3D目标体识别方法,该方法将目标体回波信号处理方法与合成孔径方法相结合,将提取的目标体信息在特征层进行了融合,然后经BP神经网络实现了分类识别,可使现有方法的不足得到显著改善。通过对3类人工靶标的实验表明,该方法可显著提高缺陷的3D识别率,能够保持在90%以上,且鲁棒性也得到明显改善。Nowadays, the main method of ultrasonic 3D target recognition was using sensors to obtain one or more one-dimensional echo in space and getting the 3D information of the target body by signal processing to realize the 3D target recognition. These methods generally exist the problem of low recognition rate and poor robustness,which restrict the popularization and application of this technology. In this paper, an ultrasonic 3D target recognition method based on visual and non-visual feature fusion was proposed. This method combined the target body echo signal processing method with the synthetic aperture method, carried out data fusion of the extracted target information in the feature layer and realized the classification recognition by the BP neural network, by which the shortage of existing methods can be significantly improved. Experiments on three kinds of artificial targets showed that the method can significantly improve the 3D recognition rate of the defect which can be kept above 90%, and the robustness was also improved obviously.

关 键 词：数据融合 超声 3D目标识别 合成孔径 可视化 非可视化 特征提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]