Due to its strong penetrability and all-day,all-weather operational capability,the passive millimeter-wave imaging technology has a broad application prospect in remote sensing and guidance.In target recognition,it needs to extract features of the initial millimeter radiometric image.According to the actual imaging characteristic of millimeter-wave,an algorithm of image feature extraction based on genetic algorithm is proposed.In this algorithm,Otsu′s method and genetic algorithm are combined to obtain the optimal segmentation threshold, and then morphology processing is implemented to the segmented image for extracting the target features.Experimental result indicates that the proposed algorithm can effectively extract the features of millimeter-wave radiometric image.