Please wait a minute...
 
国土资源遥感  2021, Vol. 33 Issue (1): 129-137    DOI: 10.6046/gtzyyg.2020067
     技术方法 本期目录 | 过刊浏览 | 高级检索 |
基于Faster R-CNN和MorphACWE模型的SAR图像高原湖泊提取
董天成1(), 杨肖1, 李卉2, 张志1(), 齐睿3
1.中国地质大学(武汉)地球物理与空间信息学院,武汉 430074
2.中国地质大学(武汉)地球科学学院,武汉 430074
3.32023部队,大连 116023
The extraction of plateau lakes from SAR images based on Faster R-CNN and MorphACWE model
DONG Tiancheng1(), YANG Xiao1, LI Hui2, ZHANG Zhi1(), QI Rui3
1. Institute of Geophysics & Geomatics, China University of Geosciences, Wuhan 430074, China
2. School of Earth Science,China University of Geosciences, Wuhan 430074, China
3. 32023 Troops, Dalian 116032, China
全文: PDF(5321 KB)   HTML  
输出: BibTeX | EndNote (RIS)      
摘要 

青藏高原湖泊是高原生态环境中最重要的自然要素之一,实现青藏高原湖泊调查与监测是现阶段迫在眉睫的任务。由于水体在SAR图像上呈现出独特的镜面反射特征,使得利用SAR图像进行湖泊的提取与分析成为当下研究热点。为进一步排除干扰地物影响、提高分类准确度,采用欧空局Sentinel-1A干涉宽幅模式的斜距单视复数产品(SLC)为主要数据源,Sentinel-2A多光谱影像Level-1C产品作为参考数据源,提出一种结合改进Faster R-CNN和MorphACWE轮廓模型的SAR图像湖泊提取算法(Faster Region-based Convolution Neural Network-MorphACWE,FR-MorphACWE)。该算法结合深度学习目标检测算法的高维特征分析和MorphACWE模型的边界提取,从综合干扰多湖泊提取角度进行分类实验评价,充分利用高原湖泊的形态学和雷达反射特征,实现西藏自治区那曲市南部至日喀则市北部高原湖泊高精度提取。实验结果表明,该算法在综合干扰多湖泊情境下准确率可达99.71%,精准率和召回率均高于98%,可作为SAR图像高原湖泊提取的新算法加以推广和应用。

服务
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章
董天成
杨肖
李卉
张志
齐睿
关键词 目标检测Faster R-CNNCV轮廓模型合成孔径雷达高原湖泊提取    
Abstract

Lakes in the Tibetan Plateau constitute one of the most important natural factors in the plateau ecological environment. So, it is an urgent task to investigate and monitor lakes in the Tibetan Plateau. Because of the unique backscatter characteristics of water body in the image, the extraction and analysis of the lake based on SAR image has become a research hotspot. In order to further eliminate the interference of surface features and improve the classification accuracy, this paper proposes a high-precision lake extraction FR-MorphACWE (Faster Region-based Convolution Neural Network-MorphACWE) model of SAR image. The Interferometric Wide Swath (IW SLC) of the European Space Agency's sentinel-1A interference wide-band mode is used as the main data source, and the sentinel-2a multispectral image level-1c product is used as the reference data source. This model combines the morphological analysis advantages of Faster R-CNN target detection algorithm and the contour extraction advantages of MorphACWE model. The classification experiments were carried out from extraction of comprehensive interference multi-lake. The target detection algorithm was applied to eliminate non - lake surface disturbance. On such a basis, the active contour model was used to extract the lake boundary, and the morphological characteristics and radar reflection characteristics of plateau lakes were fully utilized to achieve high-precision extraction of plateau lakes from the south of Naqu City to the north of Xigaze City in Tibet. The experimental results show that the accuracy of the algorithm can reach 99.71% and the accuracy and recall rate are higher than 98% in the situation of multi-lake interference.

Key wordsTarget detection    Faster R-CNN    CV model    SAR    plateau lake extraction
收稿日期: 2020-03-06      出版日期: 2021-03-18
ZTFLH:  TP79  
基金资助:中国地质调查局项目“全国矿山环境恢复治理状况遥感地质调查与监测”(DD20190705);国家自然科学基金项目“环境示踪剂辅助的高寒山区融雪径流过程模拟研究”(41401076);青海省青藏高原北部地质过程与矿产资源重点实验室开放课题共同资助(2019-KZ-01)
通讯作者: 张志
作者简介: 董天成(1996-),男,硕士研究生,研究方向为深度学习地物目标检测。Email: 741204260@qq.com
引用本文:   
董天成, 杨肖, 李卉, 张志, 齐睿. 基于Faster R-CNN和MorphACWE模型的SAR图像高原湖泊提取[J]. 国土资源遥感, 2021, 33(1): 129-137.
DONG Tiancheng, YANG Xiao, LI Hui, ZHANG Zhi, QI Rui. The extraction of plateau lakes from SAR images based on Faster R-CNN and MorphACWE model. Remote Sensing for Land & Resources, 2021, 33(1): 129-137.
链接本文:  
https://www.gtzyyg.com/CN/10.6046/gtzyyg.2020067      或      https://www.gtzyyg.com/CN/Y2021/V33/I1/129
Fig.1  研究区概况示意图
数据类型 成像时间 极化方式 空间分
辨率/m
单张图幅
大小/km2
Sentinel-1A 2019-07-01 VV 5×20 250×180
2019-07-08 VV
2019-07-13 VV
2019-07-25 VV/VH
2019-07-27 VV/VH
2019-07-31 VV
2019-08-01 VV
2019-08-06 VV
2019-08-07 VV/VH
2019-08-16 VV
Sentinel-2A 2019-07-22 10 110×110
2019-08-11
2019-09-05
2019-09-22
Tab.1  研究使用数据
Fig.2  Faster R-CNN结构图
Fig.3  改进型VGG16结构图
Fig.4  FR-MorphACWE算法架构图
Fig.5  实验流程
参数 基础学
习速率
动量 子训
练集
IOU
阈值
检测模
型训练
次数
Morph
ACWE
模型迭
代次数
取值 0.001 0.9 256 0.7 40 000 700
Tab.2  FR-MorphACWE超参数设置表
Fig.6  损失率折线图
Fig.7  多湖泊SAR图像强度值对比
Tab.3  各分类方法提取结果对比
分类方法 准确率/% 精准率/% 召回率/% F1分数 Kappa系数
OTSU阈值分割 94.612 2 92.518 4 69.244 7 79.207 3 0.761 9
模糊C均值算法 85.215 2 96.671 9 42.655 4 59.192 7 0.517 7
Mask R-CNN算法 98.791 0 94.686 3 94.429 4 94.557 7 0.938 8
FR-Morph-ACWE算法 99.716 4 98.809 7 98.636 4 98.723 0 0.985 6
Tab.4  各分类方法精度对比
Fig.8  各分类方法结果与真实值差异图
[1] 王苏民, 窦鸿身, 陈克造, 等. 中国湖泊志[M]. 北京: 科学出版社, 1998: 7.
Wang S M, Dou H S, Chen K Z, et al. China lakes[M]. Beijing: Science Press, 1998: 7.
[2] 崔雪芹. 中国科学家发起“第三极水塔计划”.[EB/OL](2019-07-13) http://news.sciencenet.cn/htmlnews/2019/7/428330.shtm.
Cui X Q. Chinese scientists launched the third pole water tower project.[EB/OL](2019-07-13) http://news.sciencenet.cn/htmlnews/2019/7/428330.shtm.
[3] 卢善龙, 肖高怀, 贾立, 等. 2000—2012年青藏高原湖泊水面时空过程数据集遥感提取[J]. 国土资源遥感, 2016,28(3):181-187.doi: 10.6046/gtzyyg.2016.03.28.
doi: 10.6046/gtzyyg.2016.03.28
Lu S L, Xiao G H, Jia L, et al. Extraction of the spatial-temporal lake water surface dataset in the Tibetan Plateau over the past 10 years[J]. Remote Sensing for Land and Resources, 2016,28(3):181-187.doi: 10.6046/gtzyyg.2016.03.28.
[4] 李德仁, 周月琴, 马洪超. 卫星雷达干涉测量原理与应用[J]. 测绘科学, 2000,25(1):9.
Li D R, Zhou Y Q, Ma H C. Principles and applications of interferometry SAR[J]. Science of Surveying and Mapping, 2000,25(1):9.
[5] 安成锦, 牛照东, 李志军, 等. 典型Otsu算法阈值比较及其SAR图像水域分割性能分析[J]. 电子与信息学报, 2010,32(9):2215-2219.
doi: 10.3724/SP.J.1146.2009.01426
An C J, Niu Z D, Li Z J, et al. Otsu threshold comparison and SAR water segmentation result analysis[J]. Journal of Electronics & Information Technology, 2010,32(9):2215-2219
[6] 李智慧, 陈秀万, 罗鹏, 等. 基于混合阈值法的Envisat影像水体提取[J]. 遥感信息, 2013,28(2):62-65.
Li Z H, Chen X W, Luo P, et al. A hybird thresholding method for water extraction in Envisat image[J]. Remote Sensing Information, 2013,28(2):62-65.
[7] Hahmann T, Wessel B. Surface water body detection in high-resolution TerraSAR-X data using active contour models[C]// Synthetic Aperture Radar (EUSAR),2010 8th European Conference on.VDE, 2010.
[8] 王庆, 廖静娟. 基于SAR数据的鄱阳湖水体提取及变化监测研究[J]. 国土资源遥感, 2010(4):94-100.doi: 10.6046/gtzyyg.2010.04.19.
Wang Q, Liao J J. Water area extraction and change detection of the Poyang Lake using SAR data[J]. Remote Sensing for Land and Resources, 2010(4):94-100.doi: 10.6046/gtzyyg.2010.04.19.
[9] Jin H, Mountrakis G, Stehman S V. Assessing integration of intensity,polarimetric scattering,interferometric coherence and spatial texture metrics in PALSAR-derived land cover classification[J]. Isprs Journal of Photogrammetry & Remote Sensing, 2014,98:70-84.
[10] 邓滢, 张红, 王超, 等. 结合纹理与极化分解的面向对象极化SAR水体提取方法[J]. 遥感技术与应用, 2016,31(4):714-723.
Deng Y, Zhang H, Wang C, et al. An object-oriented water extraction method based on texture and polarimetric decomposition feature[J]. Remote Sensing Technology and Application, 2016,31(4):714-723.
[11] 冷英, 李宁. 一种改进的变化检测方法及其在洪水监测中的应用[J]. 雷达学报, 2017,6(2):204-212.
Leng Y, Li N. Improved change detection method for flood monitoring[J]. Journal of Radars, 2017,6(2):204-212.
[12] Sghaier M O, Foucher S, Lepage R. River extraction from high-resolution SAR images combining a structural feature set and mathematical morphology[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016: 1-14.
[13] Li N, Niu S, Guo Z, et al. Dynamic waterline mapping of inland great lakes using time-series SAR data from GF-3 and S-1A satellites:A case study of DJK reservoir,China[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2019,12(11):4297-4314.
[14] Zhang Y, Chen J Q, Liu X, et al. Efficent extraction of Daihai Lake based on neural network algorithrn from high resolution SAR Image.2018 Progress in Electromagnetics Research Symposium (PIERS-Toyama),AUG 1-4,2018[C]. Toyama,JAPAN.
[15] Shang F, Hirose A. Quaternion neural-network-based PolSAR land classification in poincare-sphere-parameter space[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014,52(9):5693-5703.
[16] Yang F Y, Feng T, Xu G Y, et al. Applied method for water-body segmentation based on mask R-CNN[J]. Journal of Applied Remote Sensing, 2020,14(1), 014502.
[17] Kaiming H, Georgia G, Piotr D, et al. Mask R-CNN[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018: 1.
[18] 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 & Machine Intelligence, 2015,39(6):1137-1149.
doi: 10.1109/TPAMI.2016.2577031 pmid: 27295650
[19] Chen L C, Papandreou G, Kokkinos I, et al. DeepLab:Semantic image segmentation with deep convolutional nets,atrous convolution,and fully connected CRFs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018,40(4):834.
pmid: 28463186
[20] Chan T F, Vese L A. Active contours without edges[J]. IEEE Transactions on Image Processing, 2001,10(2):266-277.
doi: 10.1109/83.902291 pmid: 18249617
[21] Marquez-Neila P, Baumela L, Alvarez L. A Morphological approach to curvature-based evolution of curves and surfaces[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014,36(1):2-17.
doi: 10.1109/TPAMI.2013.106 pmid: 24231862
[22] 张迎春, 郭禾. 基于粗糙集和新能量公式的水平集图像分割[J]. 自动化学报, 2015,41(11):1913-1925.
Zhang Y C, Guo H. Level set image segmentation based on rough set and new energy formula[J]. Acta Automatica Sinica, 2015,41(11):1913-1925.
[23] 谢东, 龚劬, 陈小彪. 基于KFCM改进CV模型的SplitBregman图像分割方法[J]. 计算机科学, 2014(s1):153-155.
Xie D, Gong Q, Chen X B. A Split Bregman image segmentation method based on KFCM and improved CV model[J]. Computer Science, 2014(s1):153-155.
[24] 李震, 郭华东, 李新武, 等, SAR干涉测量的相干性特征分析及积雪划分[J]. 遥感学报, 2002,6(5):334-338.
Li Z, Guo H D, Li X W, et al. SAR Interferometry coherence analysis and snow mapping[J]. Journal of Remote Sensing, 2002,6(5):334-338.
[25] 马威, 陈登魁, 杨娜, 等. 时序双极化SAR开采沉陷区土壤水分估计[J]. 遥感学报, 2018,22(3):521-534.
Ma W, Chen D K, Yang N, et al. Time-series approach to estimate the soil moisture of a subsidence area by using dual polarimetric radar data[J]. Journal of Remote Sensing, 2018,22(3):521-534.
[1] 邱磊, 张学志, 郝大为. 基于深度学习的视频SAR动目标检测与跟踪算法[J]. 自然资源遥感, 2023, 35(2): 157-166.
[2] 孙盛, 蒙芝敏, 胡忠文, 余旭. 多尺度轻量化CNN在SAR图像地物分类中的应用[J]. 自然资源遥感, 2023, 35(1): 27-34.
[3] 孙宇, 黄亮, 赵俊三, 常军, 陈朋弟, 成飞飞. 结合随机擦除和YOLOv4的高空间分辨率遥感影像桥梁自动检测[J]. 自然资源遥感, 2022, 34(2): 97-104.
[4] 陈静, 陈静波, 孟瑜, 邓毓弸, 节永师, 张懿. 尺度和密度约束下基于YOLOv3的风电塔架遥感检测方法[J]. 自然资源遥感, 2021, 33(3): 54-62.
[5] 刘万军, 高健康, 曲海成, 姜文涛. 多尺度特征增强的遥感图像舰船目标检测[J]. 自然资源遥感, 2021, 33(3): 97-106.
[6] 谭熊, 王晶磊, 孙一帆. 基于特征的无人机载视频运动目标快速检测方法[J]. 国土资源遥感, 2021, 33(2): 27-32.
[7] 安健健, 孟庆岩, 胡蝶, 胡新礼, 杨健, 杨天梁. 基于Faster R-CNN的火电厂冷却塔检测及工作状态判定[J]. 国土资源遥感, 2021, 33(2): 93-99.
[8] 卫虹宇, 赵银娣, 董霁红. 基于改进RetinaNet的冷却塔目标检测[J]. 国土资源遥感, 2020, 32(4): 68-73.
[9] 江珊, 王春, 宋宏利, 刘玉锋. 基于SAR与光学遥感数据相结合的农作物种植类型识别研究[J]. 国土资源遥感, 2020, 32(4): 105-110.
[10] 周光宇, 刘邦权, 张亶. 基于变分模态分解的SAR图像目标识别方法[J]. 国土资源遥感, 2020, 32(2): 33-39.
[11] 李想, 杨灿坤, 周春平, 李小娟, 张可. 高分辨率光学卫星图像目标运动信息提取研究综述[J]. 国土资源遥感, 2019, 31(3): 1-9.
[12] 白泽朝, 汪宝存, 靳国旺, 徐青, 张红敏, 刘辉. Sentinel-1A数据矿区地表形变监测适用性分析[J]. 国土资源遥感, 2019, 31(2): 210-217.
[13] 谢奇芳, 姚国清, 张猛. 基于Faster R-CNN的高分辨率图像目标检测技术[J]. 国土资源遥感, 2019, 31(2): 38-43.
[14] 王念秦, 乔德京, 符喜优. 滤波参数对Goldstein干涉相位图滤波性能的影响分析[J]. 国土资源遥感, 2019, 31(1): 117-124.
[15] 郑鸿瑞, 徐志刚, 甘乐, 陈玲, 杨金中, 杜培军. 合成孔径雷达遥感地质应用综述[J]. 国土资源遥感, 2018, 30(2): 12-20.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
京ICP备05055290号-2
版权所有 © 2015 《自然资源遥感》编辑部
地址:北京学院路31号中国国土资源航空物探遥感中心 邮编:100083
电话:010-62060291/62060292 E-mail:zrzyyg@163.com
本系统由北京玛格泰克科技发展有限公司设计开发