可模式识别的水声成像方法

Underwater acoustic imaging method that may be used for pattern recognition

提出一种在自由环境中可以进行模式识别的水声成像方法。利用角域滤波将空间不同入射方向的平面波在阵元域中进行分离,在成像扇面内按要求设计一系列角域带通滤波器,这些滤波器的输出就是从角通带内入射的平面波在阵元域中的快拍数据,可以实现要求输入为阵元域特性数据的目标识别和底质分类等专用的模式识别算法,角域滤波器输出的平面波快拍数据的幅度作为像素值进行水声成像。采用角域滤波的方法可以进行水声成像得到所需要的声图,并且每个像素对应着相应点目标远场平面波入射到接收阵时所产生的快拍数据。仿真和实验数据处理结果表明,采用角域滤波和利用常规波束形成所得到的声图是一样的,并且两种成像方法具有相同的稳健性。

An underwater acoustic imaging method for pattern recognition in a free environment was proposed. Angular domain filtering was used to separate plane waves from different direction of space in the array domain. According to requirements in the imaging sector, a series of angular domain band-pass filters were designed, and outputs of these filters were the snapshot data from an incident plane wave within the angular pass-band in the array element domain.Some special pattern recognition algorithms, such as target recognition and bottom classification, which needed the characteristic data of the array domain, could be realized. The amplitude of snapshot data from output of angular domain filter as the pixel value could be realized underwater acoustic imaging. The method of angular domain filtering could be used for underwater acoustic imaging so obtained the desired sound image, and each pixel corresponded to a snapshot data generated by the corresponding target far-field plane wave which was incident on the receiving array.The results of simulation and experimental data processing showed that the sound image obtained by angular domain filtering was same as that used the conventional beam forming method, and two imaging methods had same robustness.