雷达与红外数据融合的近距空中目标识别

发布时间：2018-01-24 05:18

本文关键词： 目标识别贝叶斯网络数据融合小波矩特征　出处：《电光与控制》2014年09期 　论文类型：期刊论文

【摘要】：为了提高传感器目标识别性能和近距空中目标识别准确性,结合雷达和红外传感器提出了一种目标融合识别模型:对于雷达传感器,提出基于参数学习贝叶斯网络的目标识别方法,首先采用EM算法对贝叶斯网络进行参数优化,然后根据获取的目标属性信息进行目标分类;对于红外成像传感器,采用基于小波矩特征的目标识别方法,首先对目标图像进行小波矩特征提取和选择,然后通过建立的BP神经网络分类器进行目标分类;最后通过D-S证据组合法则对两部分识别结果进行融合处理,实现了基于雷达和红外数据融合的近距目标识别。仿真结果表明:和单传感器相比,所提出的模型可以更加精确地进行目标识别。
[Abstract]:In order to improve the performance of sensor target recognition and the accuracy of short-range aerial target recognition, a target fusion recognition model is proposed by combining radar and infrared sensors: for radar sensors. A method of target recognition based on parameter learning Bayesian network is proposed. First, the parameters of Bayesian network are optimized by EM algorithm, and then the target is classified according to the acquired target attribute information. For infrared imaging sensor, the target recognition method based on wavelet moment feature is adopted. Firstly, the wavelet moment feature is extracted and selected from the target image, and then the target classification is carried out by the BP neural network classifier. Finally, the recognition results of two parts are fused by D-S evidence combination rule, and the short-range target recognition based on radar and infrared data fusion is realized. The simulation results show that: compared with single sensor. The proposed model can be used for target recognition more accurately.
【作者单位】：空军工程大学航空航天工程学院;
【分类号】：TN957.52;TP18
【正文快照】： 0引言目标识别是有效了解战场态势,预测敌方意图的重要环节。随着军事科技不断发展,战场环境日趋复杂,传统的单传感器目标识别方法已难以实现目标准确识别。为了提高目标识别准确性,必须融合各传感器优势,实现多传感器目标融合识别[1-2]。雷达传感器可以准确获取笛卡尔坐标系

【相似文献】