一种基于Haar-like特征毫米波图像违禁物品自动检测方法

method based on Haar-like features for Millimeter-wave(MMW) image contraband detection is proposed to solve the problem that it is hard to auto-detect the contraband hide under the clothes by MMW imaging. Take the metal pistol detection as an example, four half-surround Haar-like featurs are designed. The rectangular eigenvalue are calculated by the integral image method, and the AdaBoost algorithm is adopted to get the strong classifier. The multi-layer classifier which formed by cascade several strong classifiers is used to detect the object. The experimental results indicate that the metal pistol in different size, different angle in MMW image can be recognized well by use this method. Moreover, it is also can be applied to other type MMW contraband detecting.