一种面向机载SAR图像的目标检测算法
|
摘要
无人机合成孔径雷达(Synthetic Aperture Radar,SAR)图像在目标检测过程中存在噪点丰富、成像相对模糊等特点,容易造成目标初检测虚警较多。特征提取和虚警去除是实现SAR图像目标准确检测的重要途径。针对目标特征提取及虚警去除,提出了一种面向机载SAR图像的目标检测算法。该算法首先进行目标初检测,再基于特征,使用球面构造神经网络模型去除虚警,实现目标的精确检测。最后,基于实拍遥感SAR影像验证了算法的有效性。
The image of Synthetic Aperture Radar( SAR) loaded on unmanned aerial vehicle( UAV) is characterized by noises and relatively blurred images,which early leads to more false alarms.The feature extraction and good false alarm removal are important factors to achieve accurate SAR image target detection.In order to implement feature extraction and false alarm removal in SAR image target detection,this paper proposes a target detection algorithm for airborne SAR images.In this algorithm,the early target detection is firstly performed,and then based on feature extraction,the spherical neural network model is used to further remove false alarms for accurate detection of targets.The test results of remote sensing SAR images taken in actual UAV show that the proposed algorithm has availability.
The image of Synthetic Aperture Radar( SAR) loaded on unmanned aerial vehicle( UAV) is characterized by noises and relatively blurred images,which early leads to more false alarms.The feature extraction and good false alarm removal are important factors to achieve accurate SAR image target detection.In order to implement feature extraction and false alarm removal in SAR image target detection,this paper proposes a target detection algorithm for airborne SAR images.In this algorithm,the early target detection is firstly performed,and then based on feature extraction,the spherical neural network model is used to further remove false alarms for accurate detection of targets.The test results of remote sensing SAR images taken in actual UAV show that the proposed algorithm has availability.
引文
[1]袁峻.无人机图像目标跟踪与定位[D].南京:南京理工大学,2017.
[2]盛文,柳健.图象纹理分析方法及其最新进展[J].无线电工程,1998,28(5):8-13.
[3]杨中悦,林伟,延伟东,等.基于热核共生矩阵的SAR图像纹理目标识别[J].西北工业大学学报,2016,34(2):356-361.
[4]曹天蕊.基于蚁群算法的目标边缘检测算法[J].山西电子技术,2018(1):6-8.
[5]宋新,罗军,王鲁平,等.一种基于小波纹理特征的目标检测方法[J].红外技术,2006,28(9):545-548.
[6]贺超.基于图像频域分析显著目标检测算法研究[D].济南:山东大学,2017.
[7]陈晓清,马君国,赵宏钟,等.基于二维小波变换的激光成像雷达目标识别算法[C]∥中国光学学会2010年光学大会论文集,2010.
[8]林青松.基于低分辨雷达回波序列轮廓像的目标分类方法研究[D].长沙:国防科学技术大学,2004.
[9]胡丹丹,姜楠.基于颜色不变量和轮廓特征的地面多圆形目标检测[J].科学技术与工程,2016,16(15):251-256.
[10] Ye W,Paulson C,Wu D,et al. A Rotation-invariant Transform for Target Detection in SAR Images[C]∥Algorithms for Synthetic Aperture Radar Imagery XV. International Society for Optics and Photonics, 2008,6970:69700W.
[11] Wolf L,Martin I. Robust Boosting for Learning from Few Examples[C]∥2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition(CVPR’05),IEEE,2005,1:359-364.
[12] Redmon J,Divvala S,Girshick R,et al. You Only Look Once:Unified,Real-Time Object Detection[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2016:779-788.
[13]王港,陈金勇,高峰,等.基于深度神经网络的遥感目标检测及特征提取[J].无线电工程,2018,48(9):760-766.
[14]刘传辉,欧剑.基于无人机图像的目标检测与跟踪技术[J].信息系统工程,2014(4):68-68.
[15] Watkins.AdaBoost中利用Haar特征进行人脸识别算法分析与总结2——级联分类器与检测过程[EB/OL].https:∥blog. csdn. net/watkinsong/article/details/7631949,2012-06-04.
[16] Watkins.AdaBoost中利用Haar特征进行人脸识别算法分析与总结1——Haar特征与积分图[EB/OL].https:∥blog.csdn.net/watkinsong/article/details/7631241,2012-06-04.
[17]张建宝,陈晓锋,刘建华.用于遥感图象分类的神经网络的构造[J].中国图象图形学报,1999,4(10):831-834.
[18]陈志祥,曹飞龙.球面神经网络的构造与逼近[J].数学学报,2012,55(4):689-700.
[19]张铃,张钹.M-P神经元模型的几何意义及其应用[J].软件学报,1998,9(5):334-338.
[2]盛文,柳健.图象纹理分析方法及其最新进展[J].无线电工程,1998,28(5):8-13.
[3]杨中悦,林伟,延伟东,等.基于热核共生矩阵的SAR图像纹理目标识别[J].西北工业大学学报,2016,34(2):356-361.
[4]曹天蕊.基于蚁群算法的目标边缘检测算法[J].山西电子技术,2018(1):6-8.
[5]宋新,罗军,王鲁平,等.一种基于小波纹理特征的目标检测方法[J].红外技术,2006,28(9):545-548.
[6]贺超.基于图像频域分析显著目标检测算法研究[D].济南:山东大学,2017.
[7]陈晓清,马君国,赵宏钟,等.基于二维小波变换的激光成像雷达目标识别算法[C]∥中国光学学会2010年光学大会论文集,2010.
[8]林青松.基于低分辨雷达回波序列轮廓像的目标分类方法研究[D].长沙:国防科学技术大学,2004.
[9]胡丹丹,姜楠.基于颜色不变量和轮廓特征的地面多圆形目标检测[J].科学技术与工程,2016,16(15):251-256.
[10] Ye W,Paulson C,Wu D,et al. A Rotation-invariant Transform for Target Detection in SAR Images[C]∥Algorithms for Synthetic Aperture Radar Imagery XV. International Society for Optics and Photonics, 2008,6970:69700W.
[11] Wolf L,Martin I. Robust Boosting for Learning from Few Examples[C]∥2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition(CVPR’05),IEEE,2005,1:359-364.
[12] Redmon J,Divvala S,Girshick R,et al. You Only Look Once:Unified,Real-Time Object Detection[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2016:779-788.
[13]王港,陈金勇,高峰,等.基于深度神经网络的遥感目标检测及特征提取[J].无线电工程,2018,48(9):760-766.
[14]刘传辉,欧剑.基于无人机图像的目标检测与跟踪技术[J].信息系统工程,2014(4):68-68.
[15] Watkins.AdaBoost中利用Haar特征进行人脸识别算法分析与总结2——级联分类器与检测过程[EB/OL].https:∥blog. csdn. net/watkinsong/article/details/7631949,2012-06-04.
[16] Watkins.AdaBoost中利用Haar特征进行人脸识别算法分析与总结1——Haar特征与积分图[EB/OL].https:∥blog.csdn.net/watkinsong/article/details/7631241,2012-06-04.
[17]张建宝,陈晓锋,刘建华.用于遥感图象分类的神经网络的构造[J].中国图象图形学报,1999,4(10):831-834.
[18]陈志祥,曹飞龙.球面神经网络的构造与逼近[J].数学学报,2012,55(4):689-700.
[19]张铃,张钹.M-P神经元模型的几何意义及其应用[J].软件学报,1998,9(5):334-338.