摘要
以信息融合的理论为基础,利用从可见光图像序列和8-12mm长波红外图像序列中提取的信息对不同种类飞机进行识别。采用矩特征并结合BP神经网络的方法,分别在特征级和决策级两个不同层次上实现了信息融合。实验结果表明,通过信息融合进行飞机识别的准确率可达到90%以上。
With information fusion theory as its basis, methods of recognizing of different kinds of aircraft by means of the information extracted from visible light image sequence and 8-12mm long wave IR image sequence is proposed . Using matrix feature and combining with BP neural network method, the information fusion is implemented in two different levels of feature level and decision level, respectively. The experimental results show that the recognition accuracy for aircraft is as high as more than 90% through information fusion.
出处
《光电工程》
CAS
CSCD
北大核心
2003年第6期50-54,72,共6页
Opto-Electronic Engineering
基金
国家自然科学基金重大项目( 60135020 F F030405)
关键词
目标识别
信息融合
特征提取
飞机识别
Target recognition
Information fusion
Feature extraction
Aircraft recognition
