Noise separation of CSEM data based on improved clustering method

In this paper, to reduce human factors and improve inefficiencies in Controlled Source Electromagnetic Method (CSEM) data processing, we present a noise separation method based on the Weighted Adaptive Bandwidth Mean-Shift Clustering (WAB-MSC) to suppress the strong electromagnetic interference in CSEM. Firstly, the kernel function is added to the traditional Mean-Shift Clustering (MSC) algorithm to reduce the sensitivity of the bandwidth selection, which can improve the robustness of the algorithm. Then, with the distribution characteristics of measured CSEM data, we proposed a bandwidth estimation method based on local density gradient to realize adaptive bandwidth selection. Finally, to verify our proposed method, the simulation signals composited by different noise types and pseudo-random signals are tested. Results show that our proposed method can effectively suppress the strong electromagnetic interference on CSEM data, which maximizes the retention of valid data and improves the signal-to-noise ratio of the data. Moreover, once the influence of strong interference noise on CSEM raw data has been effectively reduced, the apparent resistivity curves become smooth and continuous. The processing results provide a guarantee for obtaining a real geoelectric response model and subsequent data processing.