针对毫米波雷达机场跑道异物(Foreign Object Debris,FOD)检测算法存在虚警率较高,检测性能较差的问题,提出一种基于双谱特征和支撑向量数据描述(Support Vector Domain Description,SVDD)一类分类器的FOD检测方法。首先利用双谱变换将毫米波雷达接收到的FOD和杂波信号转换至差异性更大的双谱域,然后提取双谱熵和二阶统计量二维特征构成特征向量作为SVDD的输入,最后利用SVDD一类分类器在特征域实现FOD检测,同时为了提升SVDD算法性能,提出一种基于遗传模拟退火算法(Genetic Simulated Annealing Algorithm,GSAA)的参数优化方法对SVDD的核参数和惩罚因子进行全局寻优。基于77GHz毫米波雷达获取的真实机场数据开展试验,结果表明相对于传统方法所提方法不仅能够获得更高的检测性能,同时能够明显降低虚警率。
Aiming at the problems of high false alarm rate and poor detection performance of millimeter wave radar airport runway foreign body (Foreign Object Debris,FOD) detection algorithm,a FOD detection method based on bispectral features and support vector domain description (Support Vector Domain Description,SVDD) classifier is proposed.Firstly,the FOD and background clutter signals received by millimeter wave radar are transformed into bispectral domain,then the two dimensional features of bispectral entropy and second order statistics are extracted to form the feature vector as the input of SVDD.Finally,SVDD classifier is used to realize FOD detection in the feature domain.At the same time,in order to improve the performance of SVDD algorithm,a genetic simulated annealing algorithm (Genetic Simulated Annealing Algorithm,GSAA) is proposed to optimize the kernel parameters and penalty factors of SVDD.Based on the real airport data obtained by 77GHz millimeter wave radar,the experimental results show that compared with the traditional methods,the proposed method can not only obtain higher detection performance,but also significantly reduce the false alarm rate.
TANG Shuang-xia. A novel FOD detection method for millimeter wave radar[J]. LASER & INFRARED,2022,52(6):820~826