基于目标的红外特性，提出了目标的９个特征量，并以它们为分量组成目标的特征向量，以此来进行目标的分类识别．由于特征量单位不统一，提出了一种近邻分类法。文中用３类飞机的红外图像进行实验。实验表明，用这种方法把提出的红外特征量用于目标的分类识别效果很好。目标特征，图像分割，特征提取，分类识别，近邻分类法 Based on the infrared characteristics of the target, nine characteristic quantities of the target are proposed, and the target eigenvectors are composed of them as the components to classify the target. Due to the unification of feature units, a nearest neighbor classification method is proposed. The paper uses three types of aircraft infrared images of experiments. Experiments show that using this method, the proposed infrared features for the classification of the target effect is good. Target feature, image segmentation, feature extraction, classification and recognition, nearest neighbor classification