Research On Improved Back Projection Imaging Algorithm For Ground Penetrating Radar

Ground Penetrating Radar(GPR)is a nondestructive exploration technology that utilizes electromagnetic waves to detect underground targets.It inverts the information of subsurface targets based on the phenomena of reflection and refraction of electromagnetic waves at different media interfaces,and then achieves the detection and identification of targets.GPR initially interprets 1D or 2D data for target location and identification.However,this method relies on the level of expertise and engineering experience of the operator.With the GPR imaging technique,the subsurface targets’ features can be displayed intuitively,which is helpful for target interpretation.Therefore,GPR imaging possesses good development prospects.Compared with other imaging techniques,the back projection(BP)imaging algorithm has been widely used because of its simple model and strong portability.Nevertheless,the imaging method still suffers from the following two limitations: one is the slow computing speed;and the other is the high artifact energy.In order to solve the above two problems,a fast cross-correlation weighted BP imaging algorithm is proposed to improve the classical BP algorithm in the aspects of artifact suppression and imaging speed.Furthermore,the proposed algorithm is applied to the Golay complementary code GPR for underground target detection.The main contents of the thesis are as follows:(1)The classical BP imaging algorithm and several major improved BP algorithms are analyzed and summarized.Simulations and experiments are performed to compare the existing algorithms with the proposed algorithm.(2)The existing refraction point approximation solution method is improved to address the problem of time-consuming in the process of solving the refraction point position at the interface of medium.Additionally,a weighted cross-correlation BP imaging algorithm based on sliding window energy detection is proposed for suppressing the high energy artifacts,and the simulation results demonstrate the effectiveness of the method.Furthermore,the above two methods are combined to propose a fast weighted cross-correlation BP imaging algorithm,which can simultaneously improve the processing speed and artifact suppression.(3)The experimental system of Golay complementary code GPR is built,and the fast weighted cross-correlation BP algorithm is applied to this experimental system to detect single and double targets buried in dry sand.The performance of the improved algorithm is analyzed by the indicators such as target clutter ratio(TCR),peak side lobe ratio(PSLR),computational complexity and running time.Further,the improved algorithm is applied to step-frequency GPR,and its applicability is analyzed by detecting targets in different detection scenarios.