Signal Processing in Micro-Doppler Image Based Concealed Dangerous Object Detection

摘要

In this paper, we propose a signal processing method for ubiquitous concealed dangerous object detection via millimeter wave radar sensors. The proposal uses object's compound velocity image instead of optical image to make the detection. The compound velocity image is obtained through analysis of micro-Doppler signature. The proposal is extended to detect multi-persons simultaneously. A 77GHz radar sensor based experiment platform has been built to evaluate its effectiveness. Experiment results show that for 200 tests, its miss alarm rate is 3%, and detection accuracy is 93%.