红外图像中人体目标检测技术研究

针对红外序列图像中人体目标检测问题，采用了基于特征点的特征区域提取方法，先用FAST算法快速提取特征点，然后基于提取出的特征点，使用LBP算法提取特征区域，在得到感兴趣的特征区域（ROI区域）后，用对ROI区域进行基于离散小波变换的小波熵特征提取，并采用复合分类方法对ROI区域进行分类，利用此方法有效地将人体目标从红外序列图像中检测出来。

Detection technology of human targets in infrared images

A method of feature region extraction based on key-points is adopted in this paper to detect of human bodies in infrared images.FAST algorithm is used to extract the feature points rapidly first,and then the feature region is extracted with LBP algorithm according to the extracted feature points.After locating the regions of interest(ROI),the wavelet entropy feature based on discrete wavelet transform(DWT) is extracted in ROI to obtain more accurate target feature of human body.At last,the support vector machine(SVM) and Adaboost are adopted to classify ROI by the compound classification method.Experimental results show that the proposed method can detect the human targets in infrared images successfully.