Shape Estimation and Detection of Defects of a Structural Object from Acoustic Signal Using Digital Signal Processing and Intellectual Processing

我们已经开发了一种方法来估计结构对象的形状，如用于建筑的桩，从声学信号使用数字信号处理和智能处理。用小锤子敲击埋在地下的物体的头部，并在同一头部安装了拾音器，观察了反射波。通过智能处理，利用一些参数合成反射波。将合成反射波与观测反射波进行对比。通过改变参数，搜索出最相似的反射波。然后利用波动方程从参数中估计出物体的形状。我们用混凝土桩进行了大量的计算机模拟和实验，发现可以估计出桩的形状，估计误差为+10%。此外，该方法还能估计出桩缺陷的位置和大小，估计误差为+5%。

We have developed a method to estimate the shape of a structural object, such as a pile used for building, from acoustic signal using digital signal processing and intellectual processing. The head of an object, which was buried in the ground, was hit with a small hammer, and the reflection wave waas observed with a pick-up attached at the same head. By intellectual processing, the reflection wave could be synthesized using some parameters. The synthesized reflection wave and the observed reflection wave was compared with each other. By varying the parameters, the parameters ware searched, which offer the most similar reflection wave. Then, the shape of the object can be estimated from the parameters by the wave equation. We have carried out many computer simulations and experiments using concrete piles, and it is found that the shape of the pile can be estimated with an estimation error of +10%. Furthermore, the positions and the sizes of the defects of piles can be estimated with an estimation error of +5%.