基于深度学习的复杂背景雷达图像多目标检测
|
摘要
针对传统雷达图像目标检测方法在海杂波及多种干扰物组成的复杂背景下目标分类识别率低、虚警率高的问题,提出将当前热点研究的深度学习方法引入到雷达图像目标检测。首先分析了目前先进的YOLOv3检测算法优点及应用到雷达图像领域的局限,并构建了海杂波环境下有干扰物的舰船目标检测数据集,数据集包含了不同背景、分辨率、目标物位置关系等条件,能够较完备地满足实际任务需要。针对该数据集包含目标稀疏、目标尺寸小的特点,首先利用K-means算法计算适合该数据集的锚点坐标;其次在YOLOv3的基础上提出改进多尺度特征融合预测算法,融合了多层特征信息并加入空间金字塔池化。通过大量对比实验,在该数据集上,所提方法相比原YOLOv3检测精度提高了6.07%。
In the complex background of sea clutter and various interfering objects,the traditional radar image object detection method has a low detection precision and high false alarm rates.The proposed deep learning method of the current hotspot research,is introduced into radar image target detection.Firstly,the advantages of the current advanced YOLOv3 detection algorithm and the limitations applied to the radar image field are analyzed,and the ship target detection data set with interference objects in the sea clutter environment is constructed.The data set contains various backgrounds,resolution,target position relations and other conditions,which can meet the actual task needs more completely.In allusion to the target sparseness and small target size,the K-means algorithm is used to calculate the anchor coordinates suitable for the dataset.Secondly,based on YOLOv3,an improved multi-scale feature fusion prediction algorithm is proposed,which fuses the multi-layer features information and the spatial pyramid pooling.Through a large number of comparative experiments,the mAP of the proposed method is improved by 6.07%compared with the original YOLOv3.
In the complex background of sea clutter and various interfering objects,the traditional radar image object detection method has a low detection precision and high false alarm rates.The proposed deep learning method of the current hotspot research,is introduced into radar image target detection.Firstly,the advantages of the current advanced YOLOv3 detection algorithm and the limitations applied to the radar image field are analyzed,and the ship target detection data set with interference objects in the sea clutter environment is constructed.The data set contains various backgrounds,resolution,target position relations and other conditions,which can meet the actual task needs more completely.In allusion to the target sparseness and small target size,the K-means algorithm is used to calculate the anchor coordinates suitable for the dataset.Secondly,based on YOLOv3,an improved multi-scale feature fusion prediction algorithm is proposed,which fuses the multi-layer features information and the spatial pyramid pooling.Through a large number of comparative experiments,the mAP of the proposed method is improved by 6.07%compared with the original YOLOv3.
引文
[1]何友,黄勇,关键,等.海杂波中的雷达目标检测技术综述[J].现代雷达,2014,36(12):1-9.HE Y,HUANG Y,GUAN J,et al.A survey of radar target detection technology in sea clutter[J].Modern Radar,2014,36(12):1-9.
[2]WEISS M.Analysis of some modified cell-average CFAR processors in multiple target situations[J].IEEE Trans.on Aerospace and Electronic Systems,1982.
[3]ROBEY F C,FUHRMANN D R,KELLY E J,et al.A CFARadaptive matched filter detector[J].IEEE Trans.on Aerospace and Electronic Systems,1992,28(1):208-206.
[4]SHUI P L,LI D C,XU S W.Tri-featurebased detection of floating small targets in sea clutter[J].IEEE Trans.on Aerospace and Electronic Systems,2014,50(2):1416-1430.
[5]顾丹丹,许小剑.基于积分图像的快速ACCA-CFAR SAR图像目标检测算法[J].系统工程与电子技术,2014,36(2):248-253.GU D D,XU X J.Fast ACCA-FAR algorithm based on integral image for target detection from SAR images[J].Systems Engineering and Electronics,2014,36(2):248-253.
[6]HOU B,CHENG X Z,JIAO L C.Multilayer CFAR detection of ship resolution targets in very high resolution SAR images[J].IEEE Geoscience and Remote Sensing Letters,2015,12(4):811-815.
[7]刘宁波,关键,张建.基于分形可变步长LMS算法的海杂波中微弱目标检测[J].电子与信息学报,2010,32(2):371-376.LIU N B,GUAN J,ZHANG J.Low-observable target detection in sea clutter based on fractal-based variable stepsize LMS algorithm[J].Journal of Electronics&Information Technology,2010,32(2):371-376.
[8]CHEN X L,GUAN J,LIU N B,et al.Maneuvering target detection via radon-fractional Fourier translorm based long-time coherent integration[J].IEEE Trans.on Signal Processing,2014,62(4):939-953.
[9]陈唯实,李敬.基于空域特性的低空空域雷达目标检测[J].航空学报,2015,36(9):3060-3068.CHEN W S,LI J.Radar target detection in low-altitude airspace with spatial features[J].Acta Aeronautica et Astronautica Sinica,2015,36(9):3060-3068.
[10]陈唯实.基于时域特性的非相参雷达目标检测与跟踪[J].系统工程与电子技术,2016,38(8):1800-1807.CHEN W S.Target detection and tracking of non-coherent radar based on time domain characteristics[J].Systems Engineering and Electronics,2016,38(8):1800-1807.
[11]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[C]∥Proc.of the International Conference on Neural Information Processing Systems,2012:1106-1114.
[12]ZEILER M D,FERGUS R.Visualizing and understanding convolutional networks[C]∥Proc.of the European Conference on Computer Vision,2014:818-833.
[13]SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[J].Eprint Arxiv,2014.
[14]HE K,ZHANG X,REN S,et al.Deep residual learning for image recognition[C]∥Proc.of the IEEE Conference on Computer Vision and Pattern Recognition,2016:770-778.
[15]GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]∥Proc.of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR),2014.
[16]HE K M,ZHANG X Y,REN S Q,et al.Spatial pyramid pooling in deep convolutional networks for visual recognition[J].IEEETrans.on Pattern Analysis and Machine Intelligence,2015,37(9):1904-1916.
[17]GIRSHICK R.Fast R-CNN[C]∥Proc.of the IEEE International Conference on Computer Vision,2015.
[18]REN S,HE K,GIRSHICK R,et al.FasterR-CNN:towards real-time object detection with region proposal networks[C]∥Proc.of the International Conference on Neural Information Processing Systems,2015.
[19]REDMON J,DIVVALA S,GIRSHICK R,et al.You only look once:unified,real-time object detection[C]∥Proc.of the Computer Vision&Pattern Recognition,2016.
[20]REDMON J,FARHADI A.YOL09000:better,faster,stronger[C]∥Proc.of the IEEE Conference on Computer Vision and Pattern Recognition,2017:6517-6525.
[21]LIU W,ANGUELOV D,ERHAN D,et al.SSD:single shot multibox detector[C]∥Proc.of the European Conference on Computer Vision,2016:21-37.
[22]FU C,LIU W,RANGA A A,et al.DSSD:deconvolutional single shot detector[C]∥Proc.of the CoRR,2017.
[23]LIN T Y,GOYAL P,GIRSHICK R,et al.Focal loss for dense object detection[C]∥Proc.of the arXiv Preprint,2017.
[24]REDMON J,FARHADI A.YOLOv3:an incremental improvement[C]∥Proc.of the IEEE Conference on Computer Vision and Pattern Recognition,2018.
[2]WEISS M.Analysis of some modified cell-average CFAR processors in multiple target situations[J].IEEE Trans.on Aerospace and Electronic Systems,1982.
[3]ROBEY F C,FUHRMANN D R,KELLY E J,et al.A CFARadaptive matched filter detector[J].IEEE Trans.on Aerospace and Electronic Systems,1992,28(1):208-206.
[4]SHUI P L,LI D C,XU S W.Tri-featurebased detection of floating small targets in sea clutter[J].IEEE Trans.on Aerospace and Electronic Systems,2014,50(2):1416-1430.
[5]顾丹丹,许小剑.基于积分图像的快速ACCA-CFAR SAR图像目标检测算法[J].系统工程与电子技术,2014,36(2):248-253.GU D D,XU X J.Fast ACCA-FAR algorithm based on integral image for target detection from SAR images[J].Systems Engineering and Electronics,2014,36(2):248-253.
[6]HOU B,CHENG X Z,JIAO L C.Multilayer CFAR detection of ship resolution targets in very high resolution SAR images[J].IEEE Geoscience and Remote Sensing Letters,2015,12(4):811-815.
[7]刘宁波,关键,张建.基于分形可变步长LMS算法的海杂波中微弱目标检测[J].电子与信息学报,2010,32(2):371-376.LIU N B,GUAN J,ZHANG J.Low-observable target detection in sea clutter based on fractal-based variable stepsize LMS algorithm[J].Journal of Electronics&Information Technology,2010,32(2):371-376.
[8]CHEN X L,GUAN J,LIU N B,et al.Maneuvering target detection via radon-fractional Fourier translorm based long-time coherent integration[J].IEEE Trans.on Signal Processing,2014,62(4):939-953.
[9]陈唯实,李敬.基于空域特性的低空空域雷达目标检测[J].航空学报,2015,36(9):3060-3068.CHEN W S,LI J.Radar target detection in low-altitude airspace with spatial features[J].Acta Aeronautica et Astronautica Sinica,2015,36(9):3060-3068.
[10]陈唯实.基于时域特性的非相参雷达目标检测与跟踪[J].系统工程与电子技术,2016,38(8):1800-1807.CHEN W S.Target detection and tracking of non-coherent radar based on time domain characteristics[J].Systems Engineering and Electronics,2016,38(8):1800-1807.
[11]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[C]∥Proc.of the International Conference on Neural Information Processing Systems,2012:1106-1114.
[12]ZEILER M D,FERGUS R.Visualizing and understanding convolutional networks[C]∥Proc.of the European Conference on Computer Vision,2014:818-833.
[13]SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[J].Eprint Arxiv,2014.
[14]HE K,ZHANG X,REN S,et al.Deep residual learning for image recognition[C]∥Proc.of the IEEE Conference on Computer Vision and Pattern Recognition,2016:770-778.
[15]GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]∥Proc.of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR),2014.
[16]HE K M,ZHANG X Y,REN S Q,et al.Spatial pyramid pooling in deep convolutional networks for visual recognition[J].IEEETrans.on Pattern Analysis and Machine Intelligence,2015,37(9):1904-1916.
[17]GIRSHICK R.Fast R-CNN[C]∥Proc.of the IEEE International Conference on Computer Vision,2015.
[18]REN S,HE K,GIRSHICK R,et al.FasterR-CNN:towards real-time object detection with region proposal networks[C]∥Proc.of the International Conference on Neural Information Processing Systems,2015.
[19]REDMON J,DIVVALA S,GIRSHICK R,et al.You only look once:unified,real-time object detection[C]∥Proc.of the Computer Vision&Pattern Recognition,2016.
[20]REDMON J,FARHADI A.YOL09000:better,faster,stronger[C]∥Proc.of the IEEE Conference on Computer Vision and Pattern Recognition,2017:6517-6525.
[21]LIU W,ANGUELOV D,ERHAN D,et al.SSD:single shot multibox detector[C]∥Proc.of the European Conference on Computer Vision,2016:21-37.
[22]FU C,LIU W,RANGA A A,et al.DSSD:deconvolutional single shot detector[C]∥Proc.of the CoRR,2017.
[23]LIN T Y,GOYAL P,GIRSHICK R,et al.Focal loss for dense object detection[C]∥Proc.of the arXiv Preprint,2017.
[24]REDMON J,FARHADI A.YOLOv3:an incremental improvement[C]∥Proc.of the IEEE Conference on Computer Vision and Pattern Recognition,2018.