Human-body foreign-matter detection method and system based on millimetre-wave image

A human-body foreign-matter detection system based on millimeter-wave image, comprising:

• a human-body grayscale image obtaining module, configured to obtain a human body millimeter-wave grayscale image;
• a human-body contour image extraction module, configured to extract a human-body contour image from the millimeter-wave grayscale image;
• a histogram construction module, configured to construct a vertical space distribution histogram in a vertical direction and a horizontal space distribution histogram in a horizontal direction for the human-body contour image;
• a limb position obtaining module, configured to obtain positions of limbs of the human body according to the vertical space distribution histogram, the horizontal space distribution histogram and a pre-set human-body proportion model;
• a foreign-matter identification module, configured to identify foreign matters in the human-body contour image according to pre-set foreign-matter feature identification models,
• wherein when identifying foreign matters in the human-body contour image according to pre-set foreign-matter feature identification models, the foreign-matter identification module is further configured to:
• extract a grayscale gradient feature from the human-body contour image, extract an edge feature from the human-body contour image, calculate an average grayscale variance of the human-body contour image, and calculate edge smoothness of the human-body contour image;
• and determine presence or absence of the foreign matter in the human-body contour image based on the grayscale gradient feature, the edge feature, the average grayscale variance, and the edge smoothness;
• and
• a foreign-matter marking module, configured to mark a position of the foreign matter on the human-body contour image according to the positions of limbs, and to generate a foreign-matter detection result.