The target detection for GPR images based on curve fitting

摘要

Due to the complexity of the medium composition, physical characteristics and the spreading of electromagnetic waves, the extraction and automatic target detection of the Ground-penetrating Radar (GPR) is extremely complicated and has become an important issue in the practical application of GPR. In this paper, a method based on curve fitting is proposed by using the hyperbolic echo characteristics. The edge lines of GPR images firstly obtained by using image preprocessing methods and hyperbolas are fitted to approximate the edges. Finally, hyperbolic echo apex positions in the GPR images are extracted and marked on their apexes. The experimental results show that all targets are detected correctly.