用于空中红外目标检测的增强单发多框检测器方法
|
摘要
提出了一种用于空中红外目标检测的增强单发多框检测器(SSD)方法。分析了感受野与特征图层数的关系,同时采用池化和转置卷积操作的特征图双向融合机制,从整体上增强了特征的表达能力。通过引入浅层特征图的语义增强分支,并在高分辨率特征图上增加预测框,可提升小尺寸目标的定位精度。在VOC2007小目标和空中红外目标数据集上进行了对比测试,平均精度分别提高了7.1%和8.7%,此时检测速度略有下降。结果表明,增强SSD可在空中红外目标检测中获得较好的性能。
A method for enhancement of a single shot multibox detector(SSD) for aerial infrared target detection is proposed. Herein, the relationship between the sensing field and number of feature layers is analyzed, and a bidirectional feature map fusion mechanism that uses both pooling and deconvolution operations is proposed to enhance the feature expression ability. The semantic enhancement branch of the shallow feature map is introduced and the prediction boxes on the high-resolution feature map are increased, so that the positing accuracy of small-size targets is improved. Comparative experiments on the VOC2007 small object dataset and an aerial infrared target dataset reveal that the mean average precisions increase by 7.1% and 8.7%, respectively, accompanied by a slight decrease in detection speed. The results demonstrate that SSD enhancements can achieve good performance in aerial infrared target detection.
A method for enhancement of a single shot multibox detector(SSD) for aerial infrared target detection is proposed. Herein, the relationship between the sensing field and number of feature layers is analyzed, and a bidirectional feature map fusion mechanism that uses both pooling and deconvolution operations is proposed to enhance the feature expression ability. The semantic enhancement branch of the shallow feature map is introduced and the prediction boxes on the high-resolution feature map are increased, so that the positing accuracy of small-size targets is improved. Comparative experiments on the VOC2007 small object dataset and an aerial infrared target dataset reveal that the mean average precisions increase by 7.1% and 8.7%, respectively, accompanied by a slight decrease in detection speed. The results demonstrate that SSD enhancements can achieve good performance in aerial infrared target detection.
引文
[1] Erhan D,Szegedy C,Toshev A,et al.Scalable object detection using deep neural networks[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition,June 23-28,2014,Columbus,OH,USA.New York:IEEE,2014:2155-2162.
[2] Borji A,Cheng M M,Jiang H Z,et al.Salient object detection:a benchmark[J].IEEE Transactions on Image Processing,2015,24(12):5706-5722.
[3] Singla N.Motion detection based on frame difference method[J].International Journal of Information & Computation Technology,2014,4(15):1559-1565.
[4] Horn B K P,Schunck B G.Determining optical flow[J].Artificial Intelligence,1981,17(1/2/3):185-203.
[5] Barinova O,Lempitsky V,Kholi P.On detection of multiple object instances using hough transforms[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(9):1773-1784.
[6] Lowe D G.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.
[7] Dalal N,Triggs B.Histograms of oriented gradients for human detection[C]//2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR′05),June 20-25,2005,San Diego,CA,USA.New York:IEEE,2005:886-893.
[8] Burges C J C.A tutorial on support vector machines for pattern recognition[J].Data Mining and Knowledge Discovery,1998,2(2):121-167.
[9] Hastie T,Rosset S,Zhu J,et al.Multi-class AdaBoost[J].Statistics and Its Interface,2009,2(3):349-360.
[10] Ren S Q,He K M,Girshick R,et al.Faster R-CNN:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
[11] Redmon J,Farhadi A.YOLOv3:an incremental improvement[EB/OL].(2018-04-08)[2018-12-01].https://arxiv.org/abs/1804.02767.
[12] Liu W,Anguelov D,Erhan D,et al.SSD:single shot multibox dtector[M]//Leibe B,Matas J,Sebe N,et al.Computer Vision:ECCV 2016.Cham:Springer,2016,9905:21-37.
[13] Shelhamer E,Long J,Darrell T.Fully convolutional networks for semantic segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(4):640-651.
[14] He K M,Gkioxari G,Dollár P,et al.Mask R-CNN[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2018:2844175.
[15] Li Z X,Zhou F Q.FSSD:feature fusion single shot multibox detector[EB/OL].(2018-05-17)[2018-12-01].https://arxiv.org/abs/1712.00960.
[16] Cao G M,Xie X M,Yang W Z,et al.Feature-fused SSD:fast detection for small objects[J].Proceedings of SPIE,2018,10615:106151E.
[17] Fu C Y,Liu W,Ranga A,et al.DSSD:deconvolutional single shot detector[EB/OL].(2017-01-23)[2018-12-01].https://arxiv.org/abs/1701.06659.
[18] Jeong J,Park H,Kwak N.Enhancement of SSD by concatenating feature maps for object detection[EB/OL].(2017-05-26)[2018-12-01].https://arxiv.org/abs/1705.09587.
[19] Tang C,Ling Y S,Zheng K D,et al.Object detection method of multi-view SSD based on deep learning[J].Infrared and Laser Engineering,2018,47(1):126003.唐聪,凌永顺,郑科栋,等.基于深度学习的多视窗SSD目标检测方法[J].红外与激光工程,2018,47(1):126003.
[20] Simonyan K,Zisserman A.Very deep convolutional networks for large-scale image recognition[EB/OL].(2015-04-10)[2018-12-01].https://arxiv.org/abs/1409.1556.
[21] He K M,Zhang X Y,Ren S Q,et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),June 27-30,2016,Las Vegas,NV,USA.New York:IEEE,2016:770-778.
[22] Lin T Y,Goyal P,Girshick R,et al.Focal loss for dense object detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017:2999-3007.
[23] Xin P,Xu Y L,Tang H,et al.Fast airplane detection based on multi-layer feature fusion of fully convolutional networks[J].Acta Optica Sinica,2018,38(3):0315003.辛鹏,许悦雷,唐红,等.全卷积网络多层特征融合的飞机快速检测[J].光学学报,2018,38(3):0315003.
[24] Zhang Z S,Qiao S Y,Xie C H,et al.Single-shot object detection with enriched semantics[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition,June 18-23,2018,Salt Lake City,UT,USA.New York:IEEE,2018:5813-5821.
[25] Feng X Y,Mei W,Hu D S.Aerialtarget detection based on improved faster R-CNN[J].Acta Optica Sinica,2018,38(6):0615004.冯小雨,梅卫,胡大帅.基于改进Faster R-CNN的空中目标检测[J].光学学报,2018,38(6):0615004.
[26] Wang W X,Fu Y T,Dong F,et al.Infrared ship target detection method based on deep convolution neural network[J].Acta Optica Sinica,2018,38(7):0712006.王文秀,傅雨田,董峰,等.基于深度卷积神经网络的红外船只目标检测方法[J].光学学报,2018,38(7):0712006.
[2] Borji A,Cheng M M,Jiang H Z,et al.Salient object detection:a benchmark[J].IEEE Transactions on Image Processing,2015,24(12):5706-5722.
[3] Singla N.Motion detection based on frame difference method[J].International Journal of Information & Computation Technology,2014,4(15):1559-1565.
[4] Horn B K P,Schunck B G.Determining optical flow[J].Artificial Intelligence,1981,17(1/2/3):185-203.
[5] Barinova O,Lempitsky V,Kholi P.On detection of multiple object instances using hough transforms[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(9):1773-1784.
[6] Lowe D G.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.
[7] Dalal N,Triggs B.Histograms of oriented gradients for human detection[C]//2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR′05),June 20-25,2005,San Diego,CA,USA.New York:IEEE,2005:886-893.
[8] Burges C J C.A tutorial on support vector machines for pattern recognition[J].Data Mining and Knowledge Discovery,1998,2(2):121-167.
[9] Hastie T,Rosset S,Zhu J,et al.Multi-class AdaBoost[J].Statistics and Its Interface,2009,2(3):349-360.
[10] Ren S Q,He K M,Girshick R,et al.Faster R-CNN:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
[11] Redmon J,Farhadi A.YOLOv3:an incremental improvement[EB/OL].(2018-04-08)[2018-12-01].https://arxiv.org/abs/1804.02767.
[12] Liu W,Anguelov D,Erhan D,et al.SSD:single shot multibox dtector[M]//Leibe B,Matas J,Sebe N,et al.Computer Vision:ECCV 2016.Cham:Springer,2016,9905:21-37.
[13] Shelhamer E,Long J,Darrell T.Fully convolutional networks for semantic segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(4):640-651.
[14] He K M,Gkioxari G,Dollár P,et al.Mask R-CNN[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2018:2844175.
[15] Li Z X,Zhou F Q.FSSD:feature fusion single shot multibox detector[EB/OL].(2018-05-17)[2018-12-01].https://arxiv.org/abs/1712.00960.
[16] Cao G M,Xie X M,Yang W Z,et al.Feature-fused SSD:fast detection for small objects[J].Proceedings of SPIE,2018,10615:106151E.
[17] Fu C Y,Liu W,Ranga A,et al.DSSD:deconvolutional single shot detector[EB/OL].(2017-01-23)[2018-12-01].https://arxiv.org/abs/1701.06659.
[18] Jeong J,Park H,Kwak N.Enhancement of SSD by concatenating feature maps for object detection[EB/OL].(2017-05-26)[2018-12-01].https://arxiv.org/abs/1705.09587.
[19] Tang C,Ling Y S,Zheng K D,et al.Object detection method of multi-view SSD based on deep learning[J].Infrared and Laser Engineering,2018,47(1):126003.唐聪,凌永顺,郑科栋,等.基于深度学习的多视窗SSD目标检测方法[J].红外与激光工程,2018,47(1):126003.
[20] Simonyan K,Zisserman A.Very deep convolutional networks for large-scale image recognition[EB/OL].(2015-04-10)[2018-12-01].https://arxiv.org/abs/1409.1556.
[21] He K M,Zhang X Y,Ren S Q,et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),June 27-30,2016,Las Vegas,NV,USA.New York:IEEE,2016:770-778.
[22] Lin T Y,Goyal P,Girshick R,et al.Focal loss for dense object detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017:2999-3007.
[23] Xin P,Xu Y L,Tang H,et al.Fast airplane detection based on multi-layer feature fusion of fully convolutional networks[J].Acta Optica Sinica,2018,38(3):0315003.辛鹏,许悦雷,唐红,等.全卷积网络多层特征融合的飞机快速检测[J].光学学报,2018,38(3):0315003.
[24] Zhang Z S,Qiao S Y,Xie C H,et al.Single-shot object detection with enriched semantics[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition,June 18-23,2018,Salt Lake City,UT,USA.New York:IEEE,2018:5813-5821.
[25] Feng X Y,Mei W,Hu D S.Aerialtarget detection based on improved faster R-CNN[J].Acta Optica Sinica,2018,38(6):0615004.冯小雨,梅卫,胡大帅.基于改进Faster R-CNN的空中目标检测[J].光学学报,2018,38(6):0615004.
[26] Wang W X,Fu Y T,Dong F,et al.Infrared ship target detection method based on deep convolution neural network[J].Acta Optica Sinica,2018,38(7):0712006.王文秀,傅雨田,董峰,等.基于深度卷积神经网络的红外船只目标检测方法[J].光学学报,2018,38(7):0712006.