Identification and Annotation of Hidden Object in Human Terahertz Image

Terahertz (THz) detection technology is a new security technology, it is play a significant role for social public security in the current situation. In this paper, we propose a new fast recognition algorithm for detection of suspicious objects according to the characteristics of human THz images. The algorithm consists of the following steps: (1), Smoothing and using gray stretch algorithm to enhance the terahertz images, (2), Distinguish the suspicious object connected and not connected to the background images. Here, THz images are classified by using our morphological classification algorithm, (3), Extracting a full human body contour by using our Bilateral Contour Tracking Comparison algorithm(BCTC). Finally, the computer can automatically identify and mark the hidden suspicious objects in Terahertz Image. Through a large number of experiments show that the new detection algorithm accuracy is reaching 92%. Our test results show that the new algorithm is quite effective for segmentation and extraction with human body contour and less time-cost.