Abstract: Lop Nur is an important potash exploration area in China. However, the loose ground surface in the area causes low seismic excitation/reception and strong noise interference, resulting in poor-quality seismic data with low signal-to-noise ratio. To solve this problem we used two crossed seismic lines passing through well LDK01. Noise reduction technology using frequency and time dividers was adopted first to effectively suppress outlier low-frequency noise, then the cone-shaped FK filter was used to suppress the strong surface wave interference. Next, space variant surface-consistent deconvolution using varying time intervals was performed to narrow the energy difference in the received near surface excitation signals. Finally, wavelet deconvolution developed by Yu et al. was used to improve the vertical resolution. By using a combination of different high-resolution noise reduction technologies, the dynamic characteristics such as amplitude, frequency, phase and waveform were relatively maintained during data processing, and the quality of profile imaging was improved. Referencing to the geological data from well LUOXI 1, six sets of interfaces—Jurassic, Cretaceous, Paleogene, upper Paleogene, Neogene and upper Neogene bottom boundaries—have been clearly identified, which provides a basis for further improving the Lop Nur deep reservoir formation theory.

Key words: Lop Nur, potash, high resolution, noise reduction