Noise is a never-ending problem for seismic processing due to the complex acquisition environment and loopn processing pipeline. This project tries to alleviate the noise challenge by looking at the intrinsic characteristic of siemsic signal. Signal correaltion and machine-learned dictionaries are studied to find the best space for signal-noise separation of targed seismic sections or images.
Over years, I like to join in field trips for seismic station deployment. These trips give me great opportutnities to make friends from different departments as well as get close to the nature. Along the way, I developed good skills for station deployment and seismic acquisition.
Surface microseismic offers a cheaper alternative to monitor hydraulic fracturing procees but is suffering the problem of severe background noise issue. This project tries to tackle the noise change through curve-fitting techniques and results in good performance for various applications.