A proposed design for noise reduction algorithm in seismograph on high performance computing using F-K filter

This paper proposes a design of noise reduction algorithm for seismic data on a High Performance Computing (HPC) environment. The algorithm of interest is the F-k filter which can be optimized by executing it using parallel processing architecture. The implementation of the above mentioned noise filtering technique comprises of 3 stages which are Fast Fourier Transform followed by F-k filtering and finally inverse of Fast Fourier Transformation. Each of the slave nodes in the cluster would have the complete filtering algorithms and would be fed with the Seismic Shot Records in the form of binary codes. The number of Shot Records that would be given to a node at a given instance varies according to the complexity of each Shot Record. Load balancing is one of the major concerns for this project and that would be accomplished by the master node which acts as the coordinator for the entire process. © 2008 IEEE.