空间目标红外特征提取与识别技术

发布时间：2018-01-14 01:22

本文关键词：空间目标红外特征提取与识别技术　出处：《红外技术》2017年05期 　论文类型：期刊论文

【摘要】：针对空间目标红外识别中典型的类内变化大、类间变化小等问题,提出了一种用支持向量机(Support Vector Machine,SVM)分类器结合局部二值模式(Local Binary Pattern,LBP)直方图特征和灰度直方图特征的空间目标红外特征提取与识别方法。以国内某卫星和国外某卫星为研究对象,提取它们红外图像的LBP直方图特征以及灰度直方图特征;使用红外仿真软件生成两个目标在不同姿态、不同分辨率下的样本图,并分成两部分,分别作为SVM分类器的训练集和测试集。实验结果表明,LBP直方图特征和灰度直方图特征均能够以较高的准确率对空间目标进行识别,且其识别效能与目标红外图像的分辨率以及SVM核函数有关。
[Abstract]:In order to solve the problems of large change in class and small change between classes in infrared recognition of space objects, a support vector machine support Vector Machine is proposed. SVM) classifier combines local binary mode with local Binary Pattern. The infrared feature extraction and recognition method of spatial target based on LBP histogram feature and gray histogram feature. The LBP histogram features and gray histogram features of their infrared images are extracted. The infrared simulation software is used to generate the sample map of two targets with different attitude and resolution, and it is divided into two parts as the training set and test set of SVM classifier. The experimental results show that. Both the LBP histogram feature and the gray histogram feature can recognize the spatial target with high accuracy, and the recognition efficiency is related to the resolution of the infrared image and the SVM kernel function.
【作者单位】：北京卫星信息工程研究所;
【基金】：国家863计划课题(2014AA7032031E) 火箭军装备预研项目
【分类号】：TN215;TP391.41
【正文快照】： 自从1957年发射第一颗人造地球卫星以来,人类在航天领域的活动也逐渐频繁起来,随之而来的则是不断增加的在轨航天器以及空间碎片,造成空间环境日益恶化。恶劣的空间环境不仅对在轨航天器构成直接威胁[1],甚至对地面上人类的活动也形成了潜在的威胁。据不完全统计,目前在环绕地

【相似文献】