Several hours before landslides occur, part of the rock and soil body begin to burst, rub, and fracture, generating infrasound signals propagate outward. 3D advanced positioning of the landslide has remained unsolved, which is important for disaster prevention. Using the infrasound signal send from location of the landslide, this work aimed at proposing an advanced 3D matching positioning method, based on ray theory. We first proved vectority of the instantaneous sound pressure field from the perspective of air dynamics, and used ray theory to deduce a point’s instantaneous sound pressure in the air as the sum of all characteristic sound ray vectors. Based on this, the 3D vector acoustic field model was established through simulation. We also found that the triangular pyramid array has vertical and horizontal sampling capabilities, which can collect the vector signals due to the directional anisotropy between the elements. Consequently, the 4-element triangular pyramid vector array was designed to be placed in the acoustic field to receive infrasound signals. Since the infrasound signal is issued before the landslide, we then proposed the 3D matching positioning method: by calculating the cost function of the measurement signal of the triangular pyramid vector array and the copy signals calculated by the propagation model in the predetermined space, point by point, the maximum value was predicted as the infrasound source position, also as the predicted location of landslide. We performed simulations in complex mountainous terrain and applied the method to localization of infrasound sources before the occurrence of the landslide. As a result, by searching and matching in massive space points, we obtained the accurate infrasound source coordinates. This study shows the application prospects of this method for predicting geohazards position several hours in advance.

在发生滑坡之前的几个小时，部分岩石和土壤主体开始破裂，摩擦和破裂，产生次声信号向外传播。滑坡的3D高级定位仍未解决，这对于防灾至关重要。利用从滑坡位置发出的次声信号，这项工作旨在基于射线理论提出一种先进的3D匹配定位方法。我们首先从空气动力学的角度证明了瞬时声压场的矢量性，然后使用射线理论推导了空气中某个点的瞬时声压，作为所有特征声线矢量的总和。在此基础上，通过仿真建立了3D矢量声场模型。我们还发现三角形金字塔阵列具有垂直和水平采样功能，由于元件之间的方向各向异性，它可以收集矢量信号。因此，将四元素三角金字塔矢量阵列设计为放置在声场中，以接收次声信号。由于次声信号是在滑坡之前发出的，因此我们提出了3D匹配定位方法：通过计算三角形金字塔矢量阵列的测量信号和传播模型在预定空间中计算出的复制信号的成本函数，逐点最高点被预测为次声源位置，也被预测为滑坡的位置。我们在复杂的山区地形上进行了模拟，并将该方法应用于次滑源的定位，直到发生滑坡。结果是，通过在大量空间点中进行搜索和匹配，我们获得了准确的次声源坐标。这项研究显示了该方法在几小时前预测地质灾害位置的应用前景。