基于轮廓像的舰船目标识别方法研究

研究了舰船目标轮廓像形成原理,基于轮廓像的舰船目标识别方法不仅利用了雷达一维回波信息,还利用了回波序列间的相关性信息。建立了3种舰船模型,仿真获取回波序列形成的轮廓像数据,转化为二维灰度图像,提取多种特征进行描述。采用类间距离和类内距离作为依据讨论了特征选择的问题。仿真分析了信噪比、目标姿态角对识别结果的影响,给出了符合识别要求的信噪比范围以及建立模板库需要考虑的目标姿态角情况,结果表明采用轮廓像技术识别舰船目标具有良好的效果。

A Study on Ship Target Recognition Based on Outline Image

The theory of forming ship outline image is studied in this paper.Target recognition based on radar oudine image synthesizes advantages of one dimension range image and echo sequence relativity.Firstly,three ship models are founded.Oudine image data is simulated and multi-characters are extracted by transformed gray level Image.Then,the problem of character choise is studied based on best enlarged distance between different classes and reduced distance within same class.Finally,SNR effect and target aspect angle effect are discussed.The simulative recognition result proposes the appropriate SNR and samples to constitute templates. The analysis shows that radar outline image has good target recognition ability.