利用超声回波识别分层介质结构

Identification of the Structure of the Layered Media by Ultrasonic Echo

研究了利用ＢＰ神经网络进行分层介质结构超声识别的方法。应用理论模型分析了不同分层介质结构的回波，研究了用ＢＰ神经网络进行介质结构识别的算法与性能，结果表明它具有良好的识别效果，实际分层介质结构的正确识别率在９０％以上。

A new method for identifying the structure of layered media using the BP neural network is proposed.The echoes of various layered media are analyzed with the theoreticalmodels.The algorithm and the performance of the identification are studied.Satisfactory i-dentification results have been obtained with a correct identifying rate higher than 90%。