摘要
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。
A new method for extracting feature vector from infrared objects baud on 9 IR features we us proposed. The recognition of IR objects is studied using neighorhood method0,Exper- iments on three IR images used on airplanes are performed and the test results show that the sug- gested method is satisfying.
出处
《红外技术》
CSCD
北大核心
1999年第2期23-25,共3页
Infrared Technology
关键词
目标特征
图像分割
特征提取
分类识别
红外制导
Features of objects,Image Segmentation,Feature extraction,Classfied recognition, Method on neighborhood