Source location based on polarization angle information is a research hot spot in the field of shallow distributed source localization. The near-field P-wave and S-wave aliasing are severe and the polarization characteristics of the wave group are poor, making it difficult to extract the polarization angle, an important source localization parameter. To address the above difficulty, this paper proposes a method that is based on high-resolution parabolic Radon transform (HRP-Radon) and incorporates adaptive covariance matrix (ACM) to extract the polarization angle information of the wavefront. First, a data set is constructed from the data acquired by the sensor array, and this data set is morphologically corrected against the first break time information. Second, the time domain data set is converted into the Radon domain by HRP-Radon, and the direct P-wave is extracted from the aliasing information by use of the focusing characteristics of the P-wave and S-wave and the phase characteristics of the far-field P-wave. Third, a model is built using the ACM algorithm to extract the direct P-wave angle information and the angle extraction is assessed using a metric-polarization method. The numerical simulation and experimental results show that the proposed method is able to extract the wavefront angle information of the blasting near-field aliasing signal and determine the source location from the polarization angle information. This method is of practical value to the field of underground space.