摘要
基于红外多特征的目标识别技术是弹道导弹防御系统中的关键技术之一.由于在复杂场景下,单一辐射特征及其变换或者单一的温度特征及其变换都很难对所有目标进行正确的目标识别,因此需要更多特征进行目标融合识别,以此来提高识别准确性.设定典型场景,采用Adaboost算法通过红外辐射强度特征与温度特征分别对目标进行粗识别,之后在粗识别结果的基础上,调整分类器权值对目标进行精细识别,取得了较为理想的识别效果,证实了该方法的可行性与有效性.
Target recognition based on infrared multiple features is one of the key technologies in ballistic missile defense systems.Because in complex scenarios,a single radiation feature and its transformation or a single temperature feature and its transformation are difficult to achieve the correct target recognition for all targets.Therefore,multiple features are needed for target fusion identification to improve the accuracy of identification.In this paper,the typical scenario is set,the Adaboost algorithm is used to identify the target by infrared radiation intensity characteristics and temperature characteristics respectively,then,based on the coarse recognition results,the weight of the classifier is adjusted to identify the target.The recognition results are satisfactory,approving feasibility and validity of the method.
作者
戴桦宇
周玉明
黄山
尹小兵
DAI Hua-Yu;ZHOU Yu-Ming;HUANG Shan;YIN Xiao-Bing(Beijing Institute of Remote Sensing Information,Beijing 100092,China)
出处
《指挥与控制学报》
2019年第4期302-307,共6页
Journal of Command and Control
关键词
目标识别
辐射强度
特征融合
弹道导弹
trget recognition
radiation intensity
features fusion
ballistic missile
