Highway Void Disease Auto-detection Algorithm

Ground penetrating radar, used in road non-destructive detection, is a new technology developed recent years. Because of its high measure precision, it can meet the demand of measure error. It has been widely used in our country and abroad. Compared with other global physical methods, it is excellent in efficiency, information capacity, resisting disturb. What's more, it's a non-destructive technology in global physical detection.This paper focuses on researching reflected signals of GPR, and auto detection algorithm of road void. The algorithm mainly consists of two steps: first, pre-process GPR signals, which mainly transform data and remove noise component in the signals utilizing de-noising function of wavelet; Second, two methods have been researched: based on Taylor formula and curve research based on Curve Fitting, according to characters of voids. Analysing advantages and disadvantages of these two methods, this paper prefers the later.Wavelet transform and curve fitting are introduced in this paper. It researches de-noising function of wavelet transform used in GPR signal and judging function of curve fitting used in void detection. As far as wavelet transform is concerned, the paper mainly researches wavelet threshold de-noising and modulus maximum de-noising. When researching on modulus maximum de-nosing algorithm, this paper improved on the algorithm according to the character of the signals, which makes complexity of the algorithm decrease. This paper also researches least squares method, modulus maximum estimation method and subsection fitting method , which are three main methods in curve fitting. Finally this paper utilizes least squares method to judge void. The algorithm researched in the paper is satisfying via practice.