Contactless Material Identification Based on Millimeter-wave Radar and Residual Network

Material identification of objects has significant applications in industrial automation and smart homes. However, existing material recognition systems often suffer from low accuracy, high cost, or inconvenience. In this paper, we propose a contactless material identification method based on millimeter-wave radar and residual network. The method firstly utilizes a speaker to excite the micro-vibrations of the target object, and captures the radar echo data of the target through the millimeter-wave radar. By analyzing the phase variations in the reflected signals, we capture the micro-vibration characteristics of the object. After feature extraction, we construct a residual network model, which can enable high-precision material recognition. We evaluate our system through experiments on different objects of four common materials. The experimental results show that our method achieves a high recognition accuracy of 99.3%. Compared with traditional material recognition techniques, our approach has significant advantages in terms of non-destructive detection, low cost and low power consumption. Moreover, the method maintains accurate recognition even when the surface of the target is covered by plastic film or paint.