基于BBO-PSO混合优化算法的探地雷达波形反演  

Ground penetrating radar waveform inversion based on BBO-PSO hybrid optimization algorithm

作　　者：李宗玲 陈威 戴前伟[1,2] LI Zongling;CHEN Wei;DAI Qianwei(School of Geosciences and Info-Physics, Central South University, Changsha 410083, China;Key Laboratory of Metallogenic Prediction of Nonferrous Metal and Geological Environment Monitoring, Ministry of Education, Central South University, Changsha 410083, China)

机构地区：[1]中南大学地球科学与信息物理学院,长沙410083 [2]中南大学有色金属成矿预测与地质环境监测教育部重点实验室,长沙410083

出　　处：《物探化探计算技术》2020年第3期352-359,共8页Computing Techniques For Geophysical and Geochemical Exploration

基　　金：国家自然科学基金项目(41874148)。

摘　　要：探地雷达反演问题是高度非线性的,采用线性反演方法往往难以获得较好的反演效果,因此提出了将生物地理学优化算法同粒子群优化算法相结合的混合非线性反演方法。将该方法用于探地雷达时间域波形反演,采用时间域有限差分方法进行正演,以信号的均方误差函数作为目标函数,并针对波形反演的特点,在目标函数中加入波形的导数拟合差作为约束项,实现了结构层厚度和介电常数的波形反演。对比经典粒子群算法和生物地理学优化算法在多层介质仿真数据的一维波形反演中的效果,验证了该改进算法的有效性和抗噪性。Since the ground penetrating radar inversion problem is highly nonlinear,it is difficult to obtain good inversion results by linear inversion method.Therefore,a hybrid nonlinear inversion method combining biogeography-based optimization(BBO)algorithm and particle swarm optimization(PSO)algorithm is proposed.It uses the time-domain finite difference method for forward modeling,taking the mean square error of the signal as the objective function,and for the characteristics of waveform inversion.The derivative fitting difference of the waveform curve is added to the objective function as the constraint term,and the waveform inversion of the structural layer thickness and dielectric constant is realized.Comparing the effects of classical PSO and BBO algorithms in one-dimensional waveform inversion of multi-layer media simulation data,the effectiveness of the improved algorithm and its improvement in anti-noise ability are verified.

关 键 词：粒子群优化 生物地理学优化 探地雷达 波形反演 目标函数 

分 类 号：P631.4[天文地球—地质矿产勘探]