In induced-polarization (IP) surveys, the raw data are usually distorted significantly by the presence of electromagnetic (EM) interferences, including cultural noise. Several methods have been proposed to improve the signal-to-noise ratio of these data. However, signal processing in an electromagnetically noisy environment is still a challenging problem. We have determined a new and simple technique based on the analysis of the correlation between the measured potential and the injected primary current signals. This processing is applied to the data acquired using a new frequency-domain IP method called the spread-spectrum induced-polarization (SSIP) approach. In this approach, we use a pseudorandom m-sequence (also called the maximum length sequence) for the injected primary current. One of the advantages of this sequence is to be essentially spectrally flat in a given frequency range. Therefore, complex resistivity can be determined simultaneously at various frequencies. A new SSIP data set is acquired in the vicinity of Baiyin mine, Gansu Province, China. The correlation between potential difference and transmitting current signals for each period can be used to assess data quality. Only when the correlation coefficient between the two signals is greater than 0.5 can the SSIP data be used for subsequent processing and tomography. We determine what threshold value should be used for the correlation coefficient to extract high-quality apparent complex resistivity data and eliminate EM-contaminated data. We then compare the pseudosections with and without using the correlation analysis. When the correlation analysis is used, the noisy data are filtered out, and the target anomaly obtained through tomography is clearly enhanced. The inversion results of the apparent complex resistivity (amplitude and phase) for the survey area are consistent with some independent geologic and drilling information regarding the position of the ore body demonstrating the effectiveness of the approach.