Identification of Dangerous Goods in Human THZ Images

For terahertz image with low signal-to-noise ratio, serious blur and poor resolution, this paper uses the mean filter to denoise the terahertz image, and then uses Faster RCNN algorithm to detect and identify the dangerous goods in the terahertz image. Different from traditional algorithms, Faster RCNN algorithm uses traditional detection algorithms to locate, segment, extract effective features, integrate detection and recognition, and achieve automatic and rapid detection of hidden objects in the human body. The experimental results show that the proposed algorithm can effectively identify the dangerous articles of controlled knives in terahertz images, and the recognition rate can reach 89.6%.