如何利用较少的阵元个数得到比较理想的空间分辨力，准确找到噪声源位置，一直是人们比较关心的问题，因此本文介绍了基于最大似然估计的辐射噪声源近场定位方法，并利用遗传算法寻求最大似然估计的全局最优解，从而实现噪声源近场定位，其具有比常规聚焦波束形成更高的空间分辨力，且可以有效实现相干声源近场定位。通过计算机仿真详细分析了信噪比、测量距离及基阵孔径对本文算法定位性能的影响，说明了仅利用小孔径基阵就可实现辐射噪声源近场高分辨定位，最后通过湖试实验验证了本文方法的有效性，具有一定的工程应用价值。

How to use the less number of array elements to receive a good spatial resolution and find the location of noise sources accurately are always the problems that people concerned with. In this paper, it introduces near-field noise source location method based on Maximum Likelihood Estimation and the maximum likelihood estimate of the global optimal solution by using of genetic algorithm—to achieve near-field noise source location—which has higher spatial resolution comparing to conventional focused beamforming, and can achieve the coherent near-field sound source position effectively. Through computer simulation of a detailed analysis of signal to noise ratio, measuring the distance and array aperture positioning on the algorithm performance, it can be known that only by the use of small aperture array it can achieve high-resolution positioning of noise source near-field radiation. Finally, through lake test experiments it verifies the validity of the method—which has a certain engineering applications value.