计算机科学 ›› 2021, Vol. 48 ›› Issue (6A): 80-84.doi: 10.11896/jsjkx.200700185

• 图像处理&多媒体技术 • 上一篇    下一篇

基于改进DCGAN算法的遥感数据集增广方法

张曼, 李杰, 朱新忠, 沈霁, 成昊天   

  1. 上海航天电子技术研究所 上海201109
  • 出版日期:2021-06-10 发布日期:2021-06-17
  • 通讯作者: 张曼(manzhang_sh@163.com)

Augmentation Technology of Remote Sensing Dataset Based on Improved DCGAN Algorithm

ZHANG Man, LI Jie, ZHU Xin-zhong, SHEN Ji, CHENG Hao-tian   

  1. Shanghai Aerospace Electronics Technology Research Institute,Shanghai 201109,China
  • Online:2021-06-10 Published:2021-06-17
  • About author:ZHANG Man,born in 1993,master's.Her main research interests include image processing and remote sensing application.

摘要: 遥感数据集规模是深度学习目标检测算法性能的关键,如何利用少量数据生成大量标注图像成为当前的研究热点。针对这一问题,结合二次掩模技术,提出一种基于改进DCGAN算法的遥感数据集增广方法,自定义目标个数与位置,实现图像与标签的扩增,解决了基于GAN图像增广算法中无对应标签生成的问题。同时,针对DCGAN算法生成图像质量不高的问题,提出多尺度特征融合技术,优化DCGAN算法,提升图像质量。实验表明,在MNIST和PlANE两种数据集上,改进DCGAN算法生成的图像质量与图像多样性均优于DCGAN算法;在利用Tiny-YoloV2算法设计的验证实验中发现,所提算法增广的数据集,检测AP值高达85.45%,相对未增广算法与传统增广方法,AP值分别提高了16.05%和2.88%,验证了算法的有效性。

关键词: DCGAN, 多特征图, 目标检测, 数据集增广, 特征融合, 遥感图像

Abstract: The scale of the remote sensing dataset is a crucial factor to object detection algorithm's performance based on deep learning technology.How to generate large number of labeled images by using a small amount of data has become a hot research topic.To solve the problem,we propose an augmenting method of remote sensing dataset based on improved DCGAN algorithm using the secondary mask technology.In addition,the algorithmproposed in this paper can realize the amplification of images and labels by determining the number and location of targets to be generated,to solve the problem of no corresponding label generation in GAN-based image augmentation algorithm.Moreover,a multi-scale feature fusion technique for optimizing DCGAN algorithm is proposed to solve the problem of poor image quality generated by DCGAN algorithm.Experiments show that the improved DCGAN algorithm is superior to the DCGAN algorithm on both MNIST and PlANE datasets in terms of image quality and image diversity.The AP value of the dataset expanded based on the method we arranged is up to 84.45% in the design experiment using the Tiny-YoloV2 algorithm.Compared with the unaugmented algorithm and the traditional augmented method,the AP value of the method adopted in this paper is increased by 16.05% and 2.88% respectively,which fully verifies the effectiveness of the the technical scheme designed in this paper.

Key words: Augmentation of dataset, DCGAN, Feature fusion, Multi-feature map, Object detection, Remote sensing images

中图分类号: 

  • TP751
[1] LECLERC M,THARMARASA R,FLOREA M C,et al.Ship Classification Using Deep Learning Techniques for Maritime Target Tracking[C]//2018 21st International Conference on Information Fusion(FUSION).IEEE,2018:737-744.
[2] GIRSHICK R.Fast r-cnn[C]//Proceedings of the IEEE Internat Ional Conference on Computer Vision.2015:1440-1448.
[3] LIU W,ANGUELOV D,ERHAN D,et al.SSD:Single shotmultibox detector[C]//European Conference on Computer Vision.Cham:Springer,2016:21-37.
[4] REDMON J,DIVVALA S,GIRSHICK R,et al.You only look once:Unif ied,real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2016:779-788.
[5] LONG J,SHELHAMER E,DARRELL T.Fully convolutionanetworks for semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2015:3431-3440.
[6] HINTON G E,OSINDERO S,THE Y W.A fast learning algorithm for deep belief nets [J].Neural Computation,2014,18(7):1527-1554.
[7] CHATFIELD K,SIMONYAN K,VEDALDI A,et al.Return of the devil in the details:delving deep into convolutional nets [EB/OL] [2019-01-10].https://arxiv.org/pdf/1405.3531.pdf.
[8] RAITOHARJU J,RIABCHENKO E,MEISSNER K,et al.Data enrichment in fine-grained classification of aquatic macroinverte-brates [C]//Proceedings of the ICPR 2nd Workshop on Computer Vision for Analysis of Underwater Imagery.Washington,DC:IEEE Computer Society,2016:43-48.
[9] GOODFELLOW I J,POUGET-ABADIE J,MIRZA M,et al.Generative adversarial nets[C]//Proceedings of the 27th International Conference on Neural Information Processing Systems.Montreal Quebec,Canada,2014:2672-2680.
[10] RADFORD A,METZ L,CHINTALA S.Unsupervised representation learning with deep convolutional generative adversarial networks[J].arXiv:1511.06434,2015.
[11] ZEILER M D,KRISHNAN D,TAYLOR G W,et al.Deconvolutional networks[C]//2010 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).IEEE,2010:2528-2535.
[12] RATLIFFL J,BURDEN S A,SASTRYS S,et al.Characteriza-tion and Computation of Local Nash Equilibria in Continuous Games[C]//Proceedings of the 51st Annual Allerton Conference on Communication,Control,and Computing.Piscataway:IEEE Press,2013:917-924
[13] XIA G S,BAI X.DOTA:A Large-scale Dataset for Object Detection in Aerial Images[C]//IEEE Conference on Computer Vision and Pattern Recogniton.2018
[14] SZEGEDY C,LIU W,JIA Y,et al.Going deeper with convolutions [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2015:1-9.
[15] HE K,ZHANG X,REN S,et al.Deep Residual Learning forImage Recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).2016.
[1] 王坤姝, 张泽辉, 高铁杠.
基于Hachimoji DNA和QR分解的遥感图像可逆隐藏算法
Reversible Hidden Algorithm for Remote Sensing Images Based on Hachimoji DNA and QR Decomposition
计算机科学, 2022, 49(8): 127-135. https://doi.org/10.11896/jsjkx.210700216
[2] 刘冬梅, 徐洋, 吴泽彬, 刘倩, 宋斌, 韦志辉.
基于边框距离度量的增量目标检测方法
Incremental Object Detection Method Based on Border Distance Measurement
计算机科学, 2022, 49(8): 136-142. https://doi.org/10.11896/jsjkx.220100132
[3] 王灿, 刘永坚, 解庆, 马艳春.
基于软标签和样本权重优化的Anchor Free目标检测算法
Anchor Free Object Detection Algorithm Based on Soft Label and Sample Weight Optimization
计算机科学, 2022, 49(8): 157-164. https://doi.org/10.11896/jsjkx.210600240
[4] 张颖涛, 张杰, 张睿, 张文强.
全局信息引导的真实图像风格迁移
Photorealistic Style Transfer Guided by Global Information
计算机科学, 2022, 49(7): 100-105. https://doi.org/10.11896/jsjkx.210600036
[5] 程成, 降爱莲.
基于多路径特征提取的实时语义分割方法
Real-time Semantic Segmentation Method Based on Multi-path Feature Extraction
计算机科学, 2022, 49(7): 120-126. https://doi.org/10.11896/jsjkx.210500157
[6] 郁舒昊, 周辉, 叶春杨, 王太正.
SDFA:基于多特征融合的船舶轨迹聚类方法研究
SDFA:Study on Ship Trajectory Clustering Method Based on Multi-feature Fusion
计算机科学, 2022, 49(6A): 256-260. https://doi.org/10.11896/jsjkx.211100253
[7] 杨玥, 冯涛, 梁虹, 杨扬.
融合交叉注意力机制的图像任意风格迁移
Image Arbitrary Style Transfer via Criss-cross Attention
计算机科学, 2022, 49(6A): 345-352. https://doi.org/10.11896/jsjkx.210700236
[8] 祝文韬, 兰先超, 罗唤霖, 岳彬, 汪洋.
改进Faster R-CNN的光学遥感飞机目标检测
Remote Sensing Aircraft Target Detection Based on Improved Faster R-CNN
计算机科学, 2022, 49(6A): 378-383. https://doi.org/10.11896/jsjkx.210300121
[9] 马宾, 付永康, 王春鹏, 李健, 王玉立.
基于GDIoU损失函数的YOLOv4绝缘子高效定位算法
High Performance Insulators Location Scheme Based on YOLOv4 with GDIoU Loss Function
计算机科学, 2022, 49(6A): 412-417. https://doi.org/10.11896/jsjkx.210600089
[10] 陈永平, 朱建清, 谢懿, 吴含笑, 曾焕强.
基于外接圆半径差损失的实时安全帽检测算法
Real-time Helmet Detection Algorithm Based on Circumcircle Radius Difference Loss
计算机科学, 2022, 49(6A): 424-428. https://doi.org/10.11896/jsjkx.220100252
[11] 孙洁琪, 李亚峰, 张文博, 刘鹏辉.
基于离散小波变换的双域特征融合深度卷积神经网络
Dual-field Feature Fusion Deep Convolutional Neural Network Based on Discrete Wavelet Transformation
计算机科学, 2022, 49(6A): 434-440. https://doi.org/10.11896/jsjkx.210900199
[12] 蓝凌翔, 池明旻.
基于特征注意力融合网络的遥感变化检测研究
Remote Sensing Change Detection Based on Feature Fusion and Attention Network
计算机科学, 2022, 49(6): 193-198. https://doi.org/10.11896/jsjkx.210500058
[13] 陈佳舟, 赵熠波, 徐阳辉, 马骥, 金灵枫, 秦绪佳.
三维城市场景中的小物体检测
Small Object Detection in 3D Urban Scenes
计算机科学, 2022, 49(6): 238-244. https://doi.org/10.11896/jsjkx.210400174
[14] 胡伏原, 万新军, 沈鸣飞, 徐江浪, 姚睿, 陶重犇.
深度卷积神经网络图像实例分割方法研究进展
Survey Progress on Image Instance Segmentation Methods of Deep Convolutional Neural Network
计算机科学, 2022, 49(5): 10-24. https://doi.org/10.11896/jsjkx.210200038
[15] 范新南, 赵忠鑫, 严炜, 严锡君, 史朋飞.
结合注意力机制的多尺度特征融合图像去雾算法
Multi-scale Feature Fusion Image Dehazing Algorithm Combined with Attention Mechanism
计算机科学, 2022, 49(5): 50-57. https://doi.org/10.11896/jsjkx.210400093
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!