摘要
为了提高机载传感器目标识别系统的性能,提出了利用机载雷达和红外成像传感器数据融合的智能目标识别算法。对红外成像传感器,采用了基于小波矩特征和BP神经网络的目标识别算法,首先提取目标图像的小波矩特征并进行特征选择,然后通过BP神经网络对目标图像进行识别;对雷达传感器,提出了利用模糊推理的目标识别方法,首先选取适当的雷达特征,然后通过模糊推理进行识别;从雷达和红外传感器识别算法分别得到待识别目标所属类别的基本概率分配函数,用D-S证据组合规则将两个基本概率分配函数组合,最终实现了机载雷达和红外传感器的数据融合。仿真结果表明:融合后的识别效果优于单个雷达或红外传感器的识别效果。
In order to improve the performance of target recognition system,an intelligent target recognition method based on the data fusion of radar and infrared imaging sensor was proposed.For infrared sensor,the recognition algorithm of target image based on wavelet moment features and BP neural network was employed.The wavelet moment features were extracted to select some features,and the target image was recognized through BP neural network.For radar sensor,the target recognition method based on fuzzy reasoning was presented.The appropriate radar features were extracted,and the target image was recognized through fuzzy reasoning.Then,the two basic probability assignment functions were obtained from radar and infrared recognition algorithms respectively.Finally,Dempster-Shafer theory was used to combine the two basic probability assignment functions,and the data fusion of radar and infrared sensors was achieved.Simulation results indicate that the recognition performance by data fusion is much better than that by single radar sensor or single infrared sensor.
出处
《红外与激光工程》
EI
CSCD
北大核心
2010年第4期756-760,共5页
Infrared and Laser Engineering
基金
航空科学基金项目(2007ZC51038)
关键词
数据融合
目标识别
小波矩特征
神经网络
模糊推理
D-S证据理论
Data fusion
Target recognition
Wavelet moment features
Neural network
Fuzzy reasoning
Dempster-Shafer theory
