摘要
在基于模板的目标识别方法中,模板的精度是一个关键问题。由于合成孔径雷达成像机理的原因,在SAR图像中会出现迎坡缩短、顶底倒置和遮挡等现象。因此,SAR图像中的目标形状与一般光学图像中的不同。在进行模板匹配的过程中,其模板也会有相应的变化。针对该问题提出一种目标识别方法,使用新的模板生成方式,即利用OpenGL技术并根据特定目标模型生成符合SAR图像特征的目标模板,这是一种基于模型的识别方法。使用MSTAR提供的数据对该方法进行验证,试验结果证明了方法的有效性。
The accuracy of template is a key factor in the template matching.For the mechanism reason of SAR system,there are some phenomena as foreshortening,layover and shadowing.Therefore the shape of target in SAR image is different from the shape in optical image.And the shape of template also has its special character.This paper proposes a target recognition method to solve the problem.The method uses a new template generating method that generates the template accorded with SAR image character based on the target model and OpenGL technology.It is a model based target recognition method.Experimental results with MSTAR image data validate the efficiency of the proposed method.
出处
《遥测遥控》
2010年第2期51-55,共5页
Journal of Telemetry,Tracking and Command
关键词
合成孔径雷达
目标识别
模板匹配
几何畸变
Synthetic Aperture Radar
Target recognition
Template matching
Geometry aberration
