Abstract
This paper explores the property of three-axial magnetic flux leakage (MFL) signals in pipeline inspection. Then, metaheuristic optimization methods, including genetic algorithm (GA) and tabu search (TS) algorithm, are utilized to reconstruct defect profiles from three-axial MFL signals. Performances of the two methods are testified and compared, and a series of improving methods are proposed to minimize the time consumption while maintaining the accuracy of defect reconstruction. Experiments of defect reconstruction demonstrate that the proposed inversion methods have high performance in terms of both accuracy and robustness.