针对常规的人机交互方法识别探地雷达(GPR)数据变化效率低下的问题,提出一种利用高清摄像头三维(3D)人脸识别的三维探测雷达道路识别方法。 首先,给出了高清摄像头 3D 人脸识别方法框架;然后,分析利用该方法进行雷达道路识别的可行性,并利用主成分分析模型对雷达数据进行降维和特征识别;最后,在仿真实验中证明文中所提方法的可靠性和实用性,所提方法不但能够获得较高的人脸识别精度,而且可以准确识别 GPR 数据中的异常值。 文中方法不但可以将 3D 人脸识别技术应用到 GPR 识别领域,而且为道路异常识别提供了理论参考。

In view of the low efficiency of conventional human-computer interaction methods to identify ground penetrating radar (GPR) data changes, a road recognition method of three dimensional (3D) detection radar based on 3D face recognition of high definition (HD) camera is proposed. Firstly, the framework of HD camera 3D face recognition method is presented. Then, the feasibility of using this method for radar road recognition is analysed, and principal component analysis model is used to reduce dimension and feature recognition of radar data. Finally, the reliability and practicability of the proposed method are verified by simulation experiments. The proposed method can not only obtain high identification accuracy of human face, but also accurately identify outliers in GPR data. The method can not only apply 3D face recognition technology to GPR recognition field, but also provide theoretical reference for road anomaly recognition.

TN959. 3