Combination of Particle Swarm Optimization with LSSVM for Pipeline Defect Reconstruction

摘要

The nuclear function parameter and penalty parameter are pivotal factors which decide performance of Least Squares Support Vector Machines (LSSVM). Usually, most users select parameters for an LSSVM by rule of thumb, so they frequently fail to generate the optimal approaching effect for the function. In order to get optimal parameters automatically, a new approach based on particle swarm optimization and LSSVM was proposed, which automatically adjusts the parameters for LSSVM, ensuring the accuracy of parameter selection. This method was applied to pipeline 2D defect reconstruction; simulation results showed the method can overcome the difficulty of magnetic flux leakage signals, described defect geometrical characteristics, improving the reconstruction accuracy and practical value.