Application of particle swarm optimization to the estimation of the TSInSAR deformation parameter

Time-series Interferometric Synthetic Aperture Radar (TSInSAR) is a significant technology that monitors the earth's surface deformation. Obtaining a highly accurate deformation parameter is challenging because of errors caused by local extremums. In this letter, we introduce Particle Swarm Optimization (PSO) to solve this problem. The deformation rate and elevation correction parameters to be estimated can be regarded as a particle in continuous two-dimensional space. An iterative search strategy is executed so that the particles move in the solution space. Cooperation and information sharing between particles are utilized in the search strategy, which is expected to move the particles toward the best solution and avoid the local optimum. To validate the reliability of PSO, 23 Synthetic Aperture Radar (SAR) images in Beijing were exploited to retrieve deformation by PSO as well as by standard Permanent Scatterer Interferometry (PSI). We compared the results with the corresponding levelling data. Experiment suggested that there about 11% points have higher accuracy by employing PSO than those by standard PSI.