| 911 | 19 | 73 |
| 下载次数 | 被引频次 | 阅读次数 |
近几年里深度学习在图像、语音和文本等方面均取得了突破性的进展。遥感领域也开始引入深度学习作为一项重要的分析处理技术。国内在应用深度学习进行遥感影像目标检测等任务时,多数直接采用世界上效果较好的网络模型,缺乏深层次网络设计能力。通过设计遥感影像目标检测深度学习网络模型,并研究遥感影像目标在卷积神经网络下的成像特性,利用卷积特征成像和反卷积成像,研究卷积结构与网络模型特征提取的关系,为遥感领域深度学习网络模型设计提供技术支持。实验结果表明,深度神经网络用于目标检测识别具有很好的检测准确度,可用于多种目标检测;特征可视化直观展示了遥感目标特征抽象的过程。
Abstract：In recent years,the breakthrough progress in deep learning has been made in image,voice,text and other aspects.In the field of remote sensing,deep learning has also begun to be introduced as an important analytical processing technique. In domestic applications of deep learning for remote sensing image target detection and other tasks,most of them directly use one of the best network models in the world and are lack of deep-level network design capabilities.In this paper,the deep learning network model is designed for remote sensing image target detection,and the imaging characteristics of remote sensing image target under convolution neural network are studied.The relationship between convolution neural network structure and feature extraction of network model is studied by using convolution feature visualization and deconvolution visualization.It provides technical support for deep learning network model design in the field of remote sensing.The experimental results show that the deep neural network is very accurate for the detection of target and can be used to detect and identify a variety of targets. The feature visualization shows the process of abstracting the feature of remote sensing target.
[1]王鑫.基于选择性视觉注意机制的遥感图像机场检测[D].上海:复旦大学,2012:1-6.
[2]牛新,窦勇,张鹏,等.基于深度学习的光学遥感机场与飞行器目标识别技术[J].大数据,2016,1(5):54-67.
[3]LECUN Y,BOSER B,DENKER J S,et al.Backpropagation Applied to Handwritten Zip Code Recognition[J].Neural Computation,2008,1(4):541-551.
[4]GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2014:580-587.
[5]HE K,ZHANG X,REN S,et al.Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2015,37(9):1904-1916.
[6]GIRSHICK R.Fast R-CNN[C]∥IEEE International Conference on Computer Vision.IEEE,2015:1440-1448.
[7]REN S,HE K,GIRSHICK R,et al.Faster R-CNN:Towards Real-Time Object Detection with Region Proposal Networks[C]∥.Neural Information Processing Systems,2015:91-99.
[8]UIJLINGS J R R,SANDE K E A V D,GEVERS T,et al.Selective Search for Object Recognition[J].International Journal of Computer Vision,2013,104(2):154-171.
[9]ZITNICK C L,DOLLR P.Edge Boxes:Locating Object Proposals from Edges[M].Computer Vision-ECCV 2014.Springer International Publishing,2014:391-405.
[10]CHENG M M,ZHANG Z,LIN W Y,et al.BING:Binarized Normed Gradients for Objectness Estimation at 300fps[C]∥IEEE International Conference on Computer Vision.IEEE,2014:3286-3293.
[11]REDMON J,DIVVALA S,GIRSHICK R,et al.You Only Look Once:Unified,Real-Time Object Detection[C]∥IEEE International Conference on Computer Vision.IEEE,2016:779-788.
[12]DENG J,DONG W,SOCHER R,et al.Image Net:A largescale Hierarchical Image Database[C]∥IEEE International Conferenceon Computer Vision.IEEE,2009:248-255.
[13]ZEILER M D,FERGUS R.Visualizing and Understanding Convolutional Networks[C]∥Proceedingsof the European Conference on Computer Vision,2014:818-833.
[14]YOSINSKI J,CLUNE J,NGUYEN A,et al.Understanding Neural Networks Through Deep Visualization[J].ar Xiv preprint ar Xiv:1506.06579,2015.
[15]CHU B,YANG D,TADINADA R.Visualizing Residual Networks[J].ar Xiv:1701.02362v1[cs.CV],2017.
[16]ZINTGRAF L M,COHEN T S,ADEL T,et al.Visualizing Deep Neural Network Decisions:Prediction Difference Analysis[J].ar Xiv:1702.04595v1[cs.CV],2017.
[17]HE K,ZHANG X,REN S,et al.Identity Mappings in Deep Residual Networks[C]∥Proceedingsof the European Conference on Computer Vision,2016:630-645.
[18]DUMOULIN V,VISIN F.A guide to convolution arithmetic for deep learning[J].ar Xiv:1603.07285[cs.CV],2016.
[19]SPRINGENBERG J T,DOSOVITSKIY A,BROX T,et al.Striving for Simplicity:The All Convolutional Net[C]∥International Conference on Learning Representations,2015:1-14.
[20]ABADI M,BARHAM P,CHEN J,et al.Tensor Flow:a System for Large-scale Machine Learning[J].Usenix Conference on Operating Systems Design and Implementation.USENIX Association,2016,16:265-283.
基本信息:
DOI：
中图分类号:TP183;TP751
引用信息:
[1]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(09):760-766.
基金信息:
海洋公益性行业科研专项基金资助项目(201505002);; 中国电子科技集团公司航天信息应用技术重点实验室开放基金资助项目(EX166290025)