面向对象高分遥感影像典型自然地物半自动提取

doi:10.12082/dqxxkx.2021.200372

地球信息科学学报 ›› 2021, Vol. 23 ›› Issue (6): 1050-1062.doi: 10.12082/dqxxkx.2021.200372

面向对象高分遥感影像典型自然地物半自动提取

张春森¹(), 贾欣¹, 吴蓉蓉¹, 崔卫红², 史书¹, 郭丙轩³

1.西安科技大学测绘科学与技术学院,西安 710054
2.武汉大学遥感信息工程学院,武汉 430079
3.武汉大学测绘遥感信息工程国家重点实验室,武汉 430079

收稿日期:2020-07-15 修回日期:2020-10-29 出版日期:2021-06-25 发布日期:2021-08-25
作者简介:张春森（1963— ）,男,陕西西安人,博士,教授,主要从事摄影测量与高分遥感应用研究。E-mail: zhchunsen@aliyun.com
基金资助:
国家自然科学基金重大研究计划项目(92038301);自然资源部城市国土资源监测与仿真重点实验室开放基金资助课题(KF-2018-03-052)

Object Oriented Semi-automatic Extraction of Typical Natural Areas from High-Resolution Remote Sensing Images

ZHANG Chunsen¹(), JIA Xin¹, WU Rongrong¹, CUI Weihong², SHI Shu¹, GUO Bingxuan³

1. College of Geomatics, Xi'an University of Science and Technology, Xi'an 710054, China
2. School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China
3. State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing,Wuhan University,Wuhan 430079, China

Received:2020-07-15 Revised:2020-10-29 Online:2021-06-25 Published:2021-08-25
Supported by:
Major Research Plan of the National Natural Science Foundation of China(92038301);Laboratory of Urban Land Resources Monitoring and Simulation,Ministry of Natural Resources(KF-2018-03-052)

摘要/Abstract

摘要：

针对目前高空间分辨率遥感影像（简称高分遥感影像）地物全自动提取无法完全实现的现实,本文结合自然地物的光谱和纹理特征,提出一种面向对象的高分遥感影像典型自然地物半自动提取方法。首先构建最小生成树（Minimum Spanning Tree, MST）进行影像初始分割,根据影像灰度平均归一化值和标准差统计对象的光谱、纹理等特征。用户通过“种子点”交互选取提供前景样本,并基于区域邻接图（Region Adjacency Graph, RAG）寻找合并代价最小的区域扩充前景样本。在自动构建的环形缓冲区内选择背景样本,利用特征空间高斯滤波实现全连接条件随机场中均值场更新。依据全连接条件随机场描述全局信息,结合不同地物的提取准则最终得到自然地物的提取结果。以航空和高分二号(GF-2)遥感影像为实验数据,分别对林地、草地、耕地、裸地和水体等典型自然地物进行提取。结果显示,基于本文方法的航空影像典型自然地物提取总精度和Kappa值为0.959和0.948,相较于SVM方法分别提升了20.757%和0.268。高分二号(GF-2)遥感影像的提取总精度和Kappa值为0.959和0.941,相比SVM方法分别提高了1.698%和0.133。证明所给方法能够通过较少的用户交互,实现高分遥感影像典型自然地物高精度智能提取。

关键词: 高分遥感影像, 半自动提取, 面向对象, 全连接条件随机场, 最小生成树, 区域邻接图, 高斯混合模型, 支持向量机

Abstract:

Given that the automatic extraction of features from high resolution remote sensing images (HR remote sensing images) cannot be fully realized, this paper proposes an object oriented semi-automatic method to extract typical natural objects from HR remote sensing images based on spectral and textural features. We introduce the concept of "seed points" that provides limited seed points from the ground objects to be extracted to realize the auto-extraction of large-scale typical natural ground objects under the premise of ensuring the extraction accuracy. We firstly segment the initial images based on the Minimum Spanning Tree algorithm, and then the spectral and textural features of each object are calculated according to the average normalized value and standard deviation of the image gray. Secondly, users provide foreground samples through interactive selection of "seed points", and then find the regional expanded foreground samples with the least merging cost based on Region Adjacency Graph (RAG). Thirdly, the Gaussian Mixture Models (GMMs) can better adapt to the color of the image, effectively capture the subtle differences between the foreground and the background, and approximate any probability distribution with arbitrary precision. They have been widely used in image processing and model building. This paper uses GMMs to estimate whether the current object belongs to the foreground or background models. We then update the mean field in the fully connected conditional random field through the feature space Gaussian filter. The fully connected condition random field can describe the relationship between each node and all other nodes. Finally, the contour extraction results are obtained according to the extraction criteria of different natural objects and the global information description with the fully connected conditional random fields. In this study, we utilized aerial images and GF-2 images to verify our method, and extracted five typical natural objects including woodland, grassland, cropland, bare land, and waters. The results show that the total accuracy and Kappa value of typical natural features extracted from aerial images were 0.959 and 0.948, respectively. They increased by 20.757% and 0.268, respectively, compared with the SVM method. The total accuracy and Kappa value using Gaofen-2 (GF-2) remote sensing images were 0.959 and 0.941, respectively. They increased by 1.698% and 0.133, respectively, compared with SVM. Our results indicate that the proposed method can effectively extract the natural objects in HR remote sensing images with less interaction and time.

Key words: high spatial resolution remote sensing image, semi-automatic extraction, object oriented, fully connected conditional random fields, minimum spanning tree, region adjacency graph, Gaussian Mixture Models, Support Vector Machines

张春森, 贾欣, 吴蓉蓉, 崔卫红, 史书, 郭丙轩. 面向对象高分遥感影像典型自然地物半自动提取[J]. 地球信息科学学报, 2021, 23(6): 1050-1062.DOI:10.12082/dqxxkx.2021.200372

ZHANG Chunsen, JIA Xin, WU Rongrong, CUI Weihong, SHI Shu, GUO Bingxuan. Object Oriented Semi-automatic Extraction of Typical Natural Areas from High-Resolution Remote Sensing Images[J]. Journal of Geo-information Science, 2021, 23(6): 1050-1062.DOI:10.12082/dqxxkx.2021.200372

图/表 15

图1

图2

图3

图4

图5

图6

表1

图7

图8

表2

表3

图9

图10

表4

表5

参考文献 27

[1]	李雪轲, 王晋年, 张立福, 等. 面向对象的航空高光谱图像混合分类方法[J]. 地球信息科学学报, 2014,16(6):941-948. doi: 10.3724/SP.J.1047.2014.00941
	[ Li X K, Wang J N, Zhang L F, et al. A hybrid of object-based and pixel-based classification method with airborne hyperspectral imagery[J]. Journal of Geo-information Science, 2014,16(6):941-948. ]
[2]	于莉楠, 宁晓刚, 王浩, 等. 拓扑插值与光谱特征结合的复杂城区边界提取[J]. 测绘学报, 2019,48(8):1004-1013.
	[ Yu L N, Ning X G, Wang H, et al. Complex urban boundary extraction based on topological interpolation and spectral feature[J]. Acta Geodaetica et Cartographica Sinica, 2019,48(8):1004-1013. ]
[3]	程臻. 面向对象的高分辨率遥感影像全要素分类研究[D]. 哈尔滨:哈尔滨工业大学, 2015.
	[ Cheng Z. Research on object-oriented high-resolution remote sensing image classification[D]. Harbin: Harbin Institute of Technology, 2015. ]
[4]	白韬. 面向对象的GF-2遥感影像多层次分类方法研究[D]. 长春:吉林大学, 2020.
	[ Bai T. Research on object-oriented GF-2 multi-level classification of remote sensing images[D]. Changchun: Jilin University, 2020. ]
[5]	李小江, 孟庆岩, 王春梅, 等. 一种面向对象的像元级遥感图像分类方法[J]. 地球信息科学学报, 2013,15(5):744-751.
	[ Li X J, Meng Q Y, Wang C M, et al. A hybrid model of object-oriented and pixel based classification of remotely sensed data[J]. Journal of Geo-information Science, 2013,15(5):744-751. ]
[6]	姚国红, 张锦, 王励. 面向对象的高分辨率遥感影像地理要素提取[J]. 测绘科学, 2012,37(6):53-55,61.
	[ Yao G H, Zhang J, Wang L. Object-oriented feature extraction utilizing high resolution remote sensing imagery[J]. Science of Surveying and Mapping, 2012,37(6):53-55,61. ]
[7]	冯丽英. 基于深度学习技术的高分辨率遥感影像建设用地信息提取研究[D]. 杭州:浙江大学, 2017.
	[ Feng L Y. Research on information extraction of construction land from high-resolution remote sensing images based on deep learning technology[D]. Hangzhou: Zhejiang University, 2017. ]
[8]	Nguyen T N A, Cai J, Zhang J, et al. Robust interactive image segmentation using convex active contours[J]. IEEE Transactions on Image Processing, 2012,21(8):3734-3743. doi: 10.1109/TIP.2012.2191566
[9]	张煜, 陈文龙, 魏思奇, 等. 基于区域统计的遥感地物半自动提取[J]. 长江科学院院报, 2016,33(11):12-16.
	[ Zhang Y, Chen W L, Wei S Q, et al. Target extraction based on polygon statistical region-based active contour[J]. Journal of Yangtze River Scientific Research Institute, 2016,33(11):12-16. ]
[10]	王蕾, 杨武年, 任金铜, 等. GF-2影像面向对象典型城区地物提取方法[J]. 测绘通报, 2018(1):138-142.
	[ Wang L, Yang W N, Ren J T, et al. Object-oriented extraction method of typical urban features based on GF-2 images[J]. Bulletin of Surveying and Mapping, 2018(1):138-142. ]
[11]	Mayunga S D, Coleman D J, Zhang Y. Semi-automatic building extraction in dense urban settlement areas from high-resolution satellite images[J]. Survey Review, 2010,42(315):50-61. doi: 10.1179/003962609X451690
[12]	Miao Z, Wang B, Shi W, et al. A semi-automatic method for road centerline extraction from VHR images[J]. IEEE Geoscience and Remote Sensing Letters, 2014,11(11):1856-1860. doi: 10.1109/LGRS.2014.2312000
[13]	Tang M, Gorelick L, Veksler O, et al. GrabCut in one cut[J]. 2013 IEEE International Conference on Computer Vision, 2013(Iccv):1769-1776.
[14]	Cheng M M, Prisacariu V A, Zheng S, et al. Densecut: Densely connected crfs for realtime grabcut[C]// Computer Graphics Forum, 2015,34(7):193-201.
[15]	闫利, 赵展, 聂倩, 等. 利用规则进行高分辨率遥感影像地物提取[J]. 武汉大学学报·信息科学版, 2012,37(6):636-639.
	[ Yan L, Zhao Z, Nie Q, et al. Ground objects extraction from high-resolution remote sensing image with rules[J]. Geomatics and Information Science of Wuhan University, 2012,37(6):636-639. ]
[16]	刘大伟, 韩玲, 韩晓勇. 基于深度学习的高分辨率遥感影像分类研究[J]. 光学学报, 2016,36(4):306-314.
	[ Liu D W, Han L, Han X Y. Classification of high resolution remote sensing images based on deep learning[J]. Acta optica Sinica, 2016,36(4):306-314. ]
[17]	赵银娣, 张良培, 李平湘. 广义马尔可夫随机场及其在多光谱纹理影像分类中的应用[J]. 遥感学报, 2006,10(1):123-129.
	[ Zhao Y D, Zhang L P, Li P X. Universal Markov random fields and its application in multispectral textured image classification[J]. Journal of Remote Sensing, 2006,10(1):123-129. ]
[18]	谭玉敏, 槐建柱, 唐中实. 基于邻接图的面向对象遥感图像分割算法[J]. 大连海事大学学报(自然科学版), 2009,35(2):81-83.
	[ Tan Y M, Huai J Z, Tang Z S. Object-oriented adjacency graph partition algorithm for remote sensing image segmentation[J]. Journal of Dalian Maritime University, 2009,35(2):81-83. ]
[19]	巫兆聪, 胡忠文, 张谦, 等. 结合光谱、纹理与形状结构信息的遥感影像分割方法[J]. 测绘学报, 2013,42(1):44-50.
	[ Wu Z C, Hu Z W, Zhang Q, et al. On combining spectral, textural and shape features for remote sensing image segmentation[J]. Acta Geodaetica et Cartographica Sinica, 2013,42(1):44-50. ]
[20]	Cheng M M, Prisacariu V A, Zheng S, et al. Densecut: Densely connected crfs for realtime grabcut[J]. Computer Graphics Forum. 2015,34(7):193-201.
[21]	Krähenbühl P, Koltun V. Efficient inference in fully connected crfs with gaussian edge potentials[C]// Advances in neural information processing systems. 2011: 109-117.
[22]	Adams A, Baek J, Davis M A. Fast high-dimensional filtering using the permutohedral lattice[C]// Computer Graphics Forum. Oxford, UK: Blackwell Publishing Ltd, 2010,29(2):753-762.
[23]	宋晓阳, 黄耀欢, 董东林, 等. 融合数字表面模型的无人机遥感影像城市土地利用分类[J]. 地球信息科学学报, 2018,20(5):703-711. doi: 10.12082/dqxxkx.2018.170635
	[ Song X Y, Huang Y H, Dong D L, et al. Urban land use classification from UAV remote sensing images based on digital surface model[J]. Journal of Geo-information Science, 2018,20(5):703-711. ]
[24]	杨雪峰, 王雪梅. 基于决策树的多角度遥感影像分类[J]. 地球信息科学学报, 2016,18(3):416-422. doi: 10.3724/SP.J.1047.2016.00416
	[ Yang X F, Wang X M. Classification of MISR multi-angle imagery based on decision tree classifier[J]. Journal of Geo-information Science, 2016,18(3):416-422. ]
[25]	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.
	[ 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. ]
[26]	邓曾, 李丹, 柯樱海, 等. 基于改进SVM算法的高分辨率遥感影像分类[J]. 国土资源遥感, 2016,28(3):12-18.
	[ Deng Z, Li D, Ke Y H, et al. An improved SVM algorithm for high spatial resolution remote sensing image classification[J]. Remote Sensing for Land & Resources, 2016,28(3):12-18. ]
[27]	陈杰. 高分辨率遥感影像面向对象分类方法研究[D]. 长沙:中南大学, 2010.
	[ Chen J. Research on object-oriented classification method of high-resolution remote sensing image[D]. Changsha: Central South University, 2010. ]

自然地物	提取准则	具体准则描述
林地	背景选择规则区域生长规则建模规则	prb_cl_ > 0.8, prb_tex< 0.1; prb_ < 0.1 prb_cl > prbcl, prb_tex> prbtex; prb_ > prb w₁= 30, w₂= 20, w₃= 2, α = 20, β = 33, γ = 3, μ = 43
草地	背景选择规则区域生长规则建模规则	prb_cl> 0.8, prb_tex< 0.1; prb_ < 0.1 prb_cl > prbcl, prb_tex> prbtex; prb_ > prb w₁= 6, w₂= 15, w₃= 2, α = 20, β = 33, γ = 3, μ = 43
耕地	背景选择规则区域生长规则建模规则	prb_cl< 0.4, prb_tex< 0.4; prb_ <0.5 prb_cl> prbcl, prb_tex> prbtex; prb_ > prb w₁= 20, w₂= 10, w₃= 5, α = 20, β = 33, γ = 3, μ = 43
裸地、水体	背景选择规则区域生长规则建模规则	prb_ < 0.1 prb_ > 0.9 w₁= 10, w₂= 10, w₃= 2, α = 20, β = 33, γ = 3, μ = 43

地物类别	查准率P	查全率R	F1值
林地	0.916	0.886	0.901
草地	0.900	0.875	0.888
耕地	0.870	0.895	0.882
裸地	0.936	0.903	0.919
水体	0.912	0.925	0.918

	SVM方法					本文方法
	林地	草地	耕地	裸地	水体	林地	草地	耕地	裸地	水体
产品	0.633	0.449	0.956	0.252	0.985	0.946	0.951	0.969	0.974	0.958
用户	0.907	0.743	0.573	0.768	0.660	0.968	0.890	0.988	0.955	0.931
总精度	0.751					0.959
Kappa值	0.680					0.948

地物类别	查准率P	查全率R	F1值
林地	0.929	0.872	0.900
草地	0.912	0.868	0.889
耕地	0.990	0.935	0.962
裸地	0.897	0.914	0.905
水体	0.981	0.826	0.897

	SVM方法					本文方法
	林地	草地	耕地	裸地	水体	林地	草地	耕地	裸地	水体
产品	0.901	0.789	0.789	0.666	0.975	0.911	0.965	0.983	0.973	0.963
用户	0.878	0.860	0.985	0.899	0.799	0.916	0.733	0.829	0.830	0.727
总精度	0.876					0.893
Kappa值	0.808					0.941

面向对象高分遥感影像典型自然地物半自动提取

Object Oriented Semi-automatic Extraction of Typical Natural Areas from High-Resolution Remote Sensing Images

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 15

参考文献 27

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	李玉洁, 江洪, 刘宣广. 基于CatBoost的植被总初级生产力遥感模拟方法及在福建省的应用[J]. 地球信息科学学报, 2023, 25(9): 1908-1922.
[2]	林娜, 何静, 王斌, 唐菲菲, 周俊宇, 郭江. 结合植被光谱特征与Sep-UNet的城市植被信息智能提取方法[J]. 地球信息科学学报, 2023, 25(8): 1717-1729.
[3]	黄帅元, 董有福, 李海鹏. 黄土高原区SRTM1 DEM高程误差校正模型构建及对比分析[J]. 地球信息科学学报, 2023, 25(3): 669-681.
[4]	詹琪琪, 赵伟, 杨梦娇, 付浩, 李昕娟, 熊东红. 雅鲁藏布江中部流域土地沙化遥感识别[J]. 地球信息科学学报, 2022, 24(2): 391-404.
[5]	周侯伯, 肖桂荣, 林炫歆, 尹玉环. 基于特征筛选与差分进化算法优化的滑坡危险性评估方法[J]. 地球信息科学学报, 2022, 24(12): 2373-2388.
[6]	熊皓丽, 周小成, 汪小钦, 崔雅君. 基于GEE云平台的福建省10 m分辨率茶园专题空间分布制图[J]. 地球信息科学学报, 2021, 23(7): 1325-1337.
[7]	李文萍, 王伟, 高星, 伍宇明, 王学成, 刘青. 融合面向对象和分水岭算法的山地湖泊提取方法[J]. 地球信息科学学报, 2021, 23(7): 1272-1285.
[8]	邢晓语, 杨秀春, 徐斌, 金云翔, 郭剑, 陈昂, 杨东, 王平, 朱立博. 基于随机森林算法的草原地上生物量遥感估算方法研究[J]. 地球信息科学学报, 2021, 23(7): 1312-1324.
[9]	李宇君, 张磊. 基于沙地指数模型的沙地监测方法[J]. 地球信息科学学报, 2021, 23(4): 680-691.
[10]	赵泉华, 冯林达, 李玉. 基于最优极化特征组合的SAR影像湿地分类[J]. 地球信息科学学报, 2021, 23(4): 723-736.
[11]	李广洋, 寇卫利, 陈帮乾, 代飞, 强振平, 吴超. 多核学习算法及其在高光谱图像分类中的应用研究进展[J]. 地球信息科学学报, 2021, 23(3): 492-504.
[12]	陈仁丽, 吴晓青, 刘柏静, 王跃启, 高猛. 基于AIS的海上渔船捕捞活动聚集区提取方法研究[J]. 地球信息科学学报, 2021, 23(12): 2163-2173.
[13]	陈昂, 杨秀春, 徐斌, 金云翔, 张文博, 郭剑, 邢晓语, 杨东. 基于面向对象与深度学习的榆树疏林识别方法研究[J]. 地球信息科学学报, 2020, 22(9): 1897-1909.
[14]	毛亚萍, 房世峰. 基于机器学习的参考作物蒸散量估算研究[J]. 地球信息科学学报, 2020, 22(8): 1692-1701.
[15]	赵青松, 兰安军, 范泽孟, 杨青. 贵州省不同地貌形态类型土壤侵蚀强度变化的定量分析[J]. 地球信息科学学报, 2020, 22(7): 1555-1566.