为充分挖掘全极化探地雷达探测道路空洞时的数据解译与融合规律,获取更加全面、准确的探测信息,以实际道路空洞为研究对象,首先利用GprMax数值建模软件对道路空洞进行全极化探测模拟,将获取到的极化数据图像化处理后,提取原始数据中蕴藏的极化特征,再采用主成分分析法对极化图像进行特征融合,最后将融合图像上的振幅最大值与初始极化图像进行比对。结果显示:融合图像的振幅最大值在单空洞条件下较VV、HH极化分别提升了约60%, 20%,在“管线+空洞”条件下较VV极化提高了5倍,较HH极化则降低了约20%。用主成分分析法对雷达全极化数据进行融合,能得到更加真实、更高分辨率的探测图像。

In order to fully excavate the data interpretation and fusion rules when the full-polarization ground penetrating radar detects road voids and obtain more comprehensive and accurate detection information, taking the actual road cavity as the research object, this paper firstly used GprMax numerical modelling software to simulate the full-polarization detection of road voids. After image processing of the obtained polarization data, the polarization features contained in the original data were extracted, and then the feature fusion of the polarization image was carried out by principal component analysis. Finally, the maximum amplitude of the fused image was compared with the initial polarization image. The results show that the maximum amplitude of the fusion image is separately about 60% and 20% higher than that of VV and HH polarization under the condition of single cavity. It is 5 times higher than that of VV polarization under the condition of "pipeline + cavity ", and about 20% lower than HH polarization. By using principal component analysis to fuse the radar full polarization data, a more realistic and higher resolution detection image can be obtained.