Parallel Computing Based on Statistical-Cost Network-Flow Algorithm for Phase Unwrapping

We introduce parallel computing to provide high performance computing and enough computing resources for massive remote sensing data processing. Based on Condor and SNAPHU(statistical-cost, network-flow algorithm for phase unwrapping), we build a parallel computing environment in cluster system for phase unwrapping. We analyze processing results in aspects of unwrapping accuracy, memory usage and computing efficiency, and conclude influencing factors in the three aspects. The experiment shows that parallel computing effectively solve the computing problem occurred in massive remote sensing data processing, such as memory shortage and inefficient computation.