首页 > 论文 > 光学学报 > 38卷 > 8期(pp:828001--1)

三维卷积神经网络模型联合条件随机场优化的高光谱遥感影像分类

Hyperspectral Remote Sensing Image Classification Based on Three-Dimensional Convolution Neural Network Combined with Conditional Random Field Optimization

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

高光谱遥感影像分类通常基于地物光谱特征,但影像中同时还存在丰富的空间信息。空间信息的有效利用能显著提高图像分类效果。因其具有的特殊结构,卷积神经网络(CNN)已成功地应用在图像分类领域,对二维图像分类具有很好的效果。如何通过深度学习并结合空间光谱信息来提高分类性能是一个关键问题。结合高光谱影像中的空间特征与光谱信息,提出一种适合于高光谱像素级分类的深度学习三维卷积神经网络模型(3D-CNN),并在初始分类的基础上利用多标签条件随机场进行优化。选取三个通用公开高光谱数据集(Indian Pines数据集、Pavia University数据集、Pavia Center数据集)进行测试,结果表明分类优化后精度得到很大提升,总体精度可达98%,Kappa系数达到97.2%。

Abstract

Hyperspectral remote sensing image classification is usually based on the spectral features of objects, but there are plenty of spatial informations in the images. The effective use of spatial information can significantly improve the image classification effect. Because of the special structure of convolution neural network (CNN), CNN has been successfully applied in the field of image classification, and has a good effect on the classification of two-dimensional images. How to improve classification performance through deep learning combined with spatial-spectral information is a key point. Combining the spatial features and spectral information of hyperspectral images, we have developed a three-dimensional convolution neural network model (3D-CNN) for hyperspectral pixel classification, and the multi labels conditional random field is optimized on the basis of the initial classification. Three general open hyperspectral datasets (Indian Pines dataset, Pavia University dataset, Pavia Center dataset) are selected for testing. Experiments show that the accuracy is greatly improved after the classification optimization, the overall accuracy can reach 98%, and the Kappa coefficient reaches 97.2%.

Newport宣传-MKS新实验室计划
补充资料

中图分类号:TP751.1

DOI:10.3788/aos201838.0828001

所属栏目:遥感与传感器

基金项目:吉林省科技厅重点科技攻关项目(20170204034SF) 、吉林省重点科技研发项目(20180201109GX)

收稿日期:2018-01-29

修改稿日期:2018-03-09

网络出版日期:2018-04-02

作者单位    点击查看

李竺强:长光卫星技术有限公司, 吉林省卫星遥感应用技术重点实验室, 吉林 长春 130000
朱瑞飞:长光卫星技术有限公司, 吉林省卫星遥感应用技术重点实验室, 吉林 长春 130000中国科学院长春光学精密机械与物理研究所, 吉林 长春 130033
高放:长光卫星技术有限公司, 吉林省卫星遥感应用技术重点实验室, 吉林 长春 130000
孟祥玉:吉林省国土资源调查规划研究院, 吉林 长春 130061
安源:长光卫星技术有限公司, 吉林省卫星遥感应用技术重点实验室, 吉林 长春 130000
钟兴:长光卫星技术有限公司, 吉林省卫星遥感应用技术重点实验室, 吉林 长春 130000

联系人作者:李竺强(skybelongtous@foxmail.com)

【1】Baraniuk R G. More is less: signal processing and the data deluge[J]. Science, 2011, 331(6018): 717-719.

【2】Li D R, Zhang L P, Xia G S. Automatic analysis and mining of remote sensing big data[J]. Acta Geodaetica et Cartographica Sinica, 2014, 43(12): 1211-1216.
李德仁, 张良培, 夏桂松. 遥感大数据自动分析与数据挖掘[J]. 测绘学报, 2014, 43(12): 1211-1216.

【3】Sun L, Wu Z B, Liu J J, et al. Supervised spectral-spatial hyperspectral image classification with weighted Markov random fields[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(3): 1490-1503.

【4】Du P J, Xia J S, Xue Z H, et al. Review of hyperspectral remote sensing image classification[J]. Journal of Remote Sensing, 2016, 20(2): 236-256.
杜培军, 夏俊士, 薛朝辉, 等. 高光谱遥感影像分类研究进展[J]. 遥感学报, 2016, 20(2): 236-256.

【5】Lacar F M, Lewis M M, Grierson I T. Use of hyperspectral imagery for mapping grape varieties in the Barossa valley[J]. Proceedings of the IEEE, 2001, 6: 2875-2877.

【6】Plaza A, Du Q, Chang Y L, et al. High performance computing for hyperspectral remote sensing[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2011, 4(3): 528-544.

【7】Samaniego L, Bardossy A, Schulz K. Supervised classification of remotely sensed imagery using a modified κ-NN technique[J]. IEEE Transactions on Geoscience and Remote Sensing, 2008, 46(7): 2112-2125.

【8】Ediriwickrema J, Khorram S. Hierarchical maximum-likelihood classification for improved accuracies[J]. IEEE Transactions on Geoscience and Remote Sensing, 1997, 35(4): 810-816.

【9】Li L, Dong Z L. Color image segmentation using improved graph cuts[J]. Geomatics and Information Science of Wuhan University, 2014, 39(12): 1504-1508.
李磊, 董卓莉. 利用改进图割的彩色图像分割算法[J]. 武汉大学学报(信息科学版), 2014, 39(12): 1504-1508.

引用该论文

Li Zhuqiang,Zhu Ruifei,Gao Fang,Meng Xiangyu,An Yuan,Zhong Xing. Hyperspectral Remote Sensing Image Classification Based on Three-Dimensional Convolution Neural Network Combined with Conditional Random Field Optimization[J]. Acta Optica Sinica, 2018, 38(8): 0828001

李竺强,朱瑞飞,高放,孟祥玉,安源,钟兴. 三维卷积神经网络模型联合条件随机场优化的高光谱遥感影像分类[J]. 光学学报, 2018, 38(8): 0828001

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF