Please wait a minute...
img

官方微信

遥感技术与应用  2021, Vol. 36 Issue (2): 265-274    DOI: 10.11873/j.issn.1004-0323.2021.2.0265
CNN 专栏     
基于卷积神经网络的面向对象露天采场提取
胡乃勋1(),陈涛1,3(),甄娜2,牛瑞卿1
1.中国地质大学(武汉)地球物理与空间信息学院,湖北 武汉 430074
2.河南省地质环境监测院,河南 郑州 450006
3.青海省地理空间信息技术与应用重点实验室,青海 西宁 810001
Object-oriented Open Pit Extraction based on Convolutional Neural Network
Naixun Hu1(),Tao Chen1,3(),Na Zhen2,Ruiqing Niu1
1.Institute of Geophysics and Geomatics,China University of Geosciences,Wuhan 430074,China
2.Geological Environment Monitoring Institute of Henan Province,Zhengzhou 450006,China
3.Geomatics Technology and Application key Laboratory of Qinghai Province,Xining 810001,China
 全文: PDF(25850 KB)   HTML
摘要:

矿产资源的过度开发会对自然环境造成严重的负影响,矿山环境监测对生态文明建设具有十分重要意义。在目前的矿山环境监测中,机器学习算法被广泛的使用并取得了较为良好的效果。近年来,随着深度学习领域的快速发展,相关理论知识也逐渐被应用于遥感图像处理中。将深度学习算法与面向对象的思想相结合,以高分二号影像作为研究数据,使用卷积神经网络对河南省禹州市的采矿区进行了以露天采场为主的开发占地类型信息提取,并与支持向量机方法进行对比,最终得到卷积神经网络的总体精度为91.85%,Kappa系数为0.90,均高于支持向量机方法,提取结果也与实际更加相符。表明该方法在露天采场提取中的优势和可行性,可为矿区的环境监测和科学管理提供可靠的技术支撑。

关键词: 露天采场面向对象卷积神经网络深度学习矿产资源    
Abstract:

The overexploitation of mineral resources will have a serious negative impact on the natural environment. The monitoring of the mine environment is of great significance to the construction of ecological civilization. Machine learning algorithms have been widely used in traditional mine monitoring and have achieved good results. In recent years, with the rapid development of the field of deep learning, relevant theoretical knowledge has gradually been applied to remote sensing image processing. In this study, the deep learning algorithm is combined with the object-oriented method, and the GF-2 image is used to extract the land occupation type by applying the conventional neural network from the mining area in Yuzhou City, Henan Province. To compare the performance of the proposed methods, the support vector machine method was used. The results show that the overall accuracy of the convolutional neural network is 91.85% and the kappa coefficient is 0.90, which is higher than the support vector machine method. This paper shows the advantages and feasibility of this method in the extraction of open-pit mining areas and provides reliable technical support for the scientific management and environmental monitoring of open-pit mining areas.

Key words: Open Pit    Object oriented    Convolutional Neural network    Deep learning    Mineral resources
收稿日期: 2019-12-14 出版日期: 2021-05-24
ZTFLH:  X87  
基金资助: 国家自然科学基金项目(62071439)
通讯作者: 陈涛     E-mail: hnx0908@cug.edu.cn;taochen@cug.edu.cn
作者简介: 胡乃勋(1993-),男,黑龙江佳木斯人,硕士研究生,主要从事高分辨率遥感地学应用研究。Email:hnx0908@cug.edu.cn
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  
胡乃勋
陈涛
甄娜
牛瑞卿

引用本文:

胡乃勋,陈涛,甄娜,牛瑞卿. 基于卷积神经网络的面向对象露天采场提取[J]. 遥感技术与应用, 2021, 36(2): 265-274.

Naixun Hu,Tao Chen,Na Zhen,Ruiqing Niu. Object-oriented Open Pit Extraction based on Convolutional Neural Network. Remote Sensing Technology and Application, 2021, 36(2): 265-274.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2021.2.0265        http://www.rsta.ac.cn/CN/Y2021/V36/I2/265

图1  研究区位置图
图2  露天采场解译标志
图3  露天采场现场照片
图4  技术流程图
图5  卷积层原理图
图6  CNN结构图
分割对象分割尺度形状/光谱权值紧致度/平滑度权值
水体、植被1500.1/0.90.5/0.5
露天采场、矿山堆积、裸土900.4/0.60.5/0.5
道路、建筑500.4/0.60.5/0.5
表1  分割尺度和权重值
图7  局部分割图
对象特征域特征
光谱特征标准差、均值、比率、亮度
纹理特征能量、熵、惯性矩、相关系数、均值
几何特征形状指数、圆度、长宽比、紧致度、主方向、非对称性、密度
表2  影像对象的特征选择
图8  原始训练过程曲线
图9  改进后训练过程曲线
图10  CNN分类结果
图11  SVM分类结果
图12  分类结果对比
露天采场道路水体植被建筑物矿山堆积裸土小计
露天采场1 948/1 82221/245/380/047/9685/1031/12 107/2 084
道路25/93911/8551/00/230/908/202/5977/1 065
水体15/40/0117/780/00/11/00/0133
植被3/40/20/2991/9903/190/44/111 001/1 032
建筑42/11516/820/52/11 873/1 7764/292/11 939/2 009
矿山堆积149/15636/210/10/244/16489/43612/27730/659
裸土17/53/31/05/33/29/4393/369431/386
小计2 199/2 199987/987124/124998/9982 000/2 000596/596414/4147 318/7 318
表3  CNN/SVM分类结果精度评价混淆矩阵
OAKappaUAPA
CNN91.86%0.9088.59%92.45%
SVM86.44%0.8382.86%87.43%
表4  CNN与SVM露天采场分类结果精度评价
图13  扩展研究区影像
图14  扩展研究区分类结果
1 Braghina C, Peptenatu D, Constantinescu S, et al. The Pressure Exerted on the Natural Environment in the Open Pit Exploitation Areas in Oltenia[J]. Carpathian Journal of Earth and Environmental Sciences, 2010,5(1):33-40.
2 Chen Weitao, Zhang Zhi, Wang Yanxin. Advances in Remote Sensing-based Detecting of Mine Exploitation and Mine Environment[J]. Remote Sensing for Land & Resource,2009,21(2):1-8.
2 陈伟涛, 张志, 王焰新. 矿山开发及矿山环境遥感探测研究进展[J]. 国土资源遥感,2009,21(2):1-8.
3 Yu Kai, Jia Lei, Chen Yuqiang, et al. Deep Learning: Yesterday, Today, and Tomorrow[J]. Journal of Computer Research and Development, 2013,50(9):1799-1804.
3 余凯, 贾磊, 陈雨强, 等. 深度学习的昨天、今天和明天[J]. 计算机研究与发展, 2013,50(9):1799-1804.
4 He Qing, Li Ning, Luo WenJuan, et al. A Survey of Machine Learning Algorithms for Big Data[J]. Pattern Recognition and Artificial Intelligence, 2014,27(4): 327-336.
4 何清, 李宁, 罗文娟, 等. 大数据下的机器学习算法综述[J]. 模式识别与人工智能, 2014,27(4):327-336.
5 Zhang Xin. The Study of SVM-based Medical Image Classification[D]. Hangzhou: Zhejiang University,2006.
5 张昕. 基于SVM方法的医学图像分类研究[D]. 杭州:浙江大学,2006.
6 Wang Shuqin. Machine Learning Methods and Their Applications in Bioinformatics[D]. Changchun: Jilin University,2017.
6 王淑琴. 机器学习方法及其在生物信息学领域中的应用[D].长春:吉林大学,2009.
7 Chen Huali, Chen Zhihua, Ding Guoping. The Application of The Knowledge-based Decision Tree Classification Method to The Extraction of Land Types in Mining Areas: A Case Study of Daye Area, Hubei Province[J]. Remote Sensing for Land & Resource, 2004,16(3): 49-53.
7 陈华丽, 陈植华, 丁国平. 用基于知识的决策树方法分层提取矿区土地类型——以湖北大冶为例[J].国土资源遥感, 2004,16(3):49-53.
8 Li Deren, Tong Qingxi, Li Rongxing, et al. Several Cutting-edge Scientific Issues of High-resolution Earth Observation[J]. ScientiaSinica(Terrae),2012,42(6):805-813.
8 李德仁, 童庆禧, 李荣兴, 等. 高分辨率对地观测的若干前沿科学问题[J]. 中国科学:地球科学, 2012,42(6):805-813.
9 Baatz M, Schape A. Multiresolution Segmentation: An Optimization Approach for High Quality Multi-scale Image Segmentation[C]∥ Angewandte Geographische Informationsverarbeitung XII Berlin, 2000:12-23.
10 Cheng Lu. Application of Object-oriented Combined SVM in Information Extraction of Open-pit Mine[D]. Xining: Qinghai University, 2017.
10 程璐. 面向对象结合支持向量机(SVM)在露天矿区信息提取中的应用研究[D]. 西宁:青海大学,2012.
11 Krizhevsky A, Sutskever I, Hinton G E. ImageNet Classification with Deep Convolutional Neural Networks[J]. Communications of the Acm, 2017,60(6):84-90.
12 Lecun Y, Jackel L D, Boser B, et al. Handwritten Digit Recognition-Applications of Neural Network Chips and Automatic Learning[J]. IEEE Communications Magazine, 1989,27(11):41-46.
13 Simonyan K, Zisserman A. Very Deep Convolutional Networks for Large-scale Image Recognition[J]. Computer Science, 2014,arXiv preprint arXiv:1409-1556.
14 Szegedy C, Liu W, Jia Y, et al. Going Deeper with Convolutions[C]∥ Proccedings of the IEEE Conference on Computer Vision and Patten Recogntion[C]∥ 2015: 1-9.
15 Xu Gang, Yue Jiguang, Dong Yanchao, et al. Cement Plant Detection on Satellite Images Using Deep Convolution Network[J]. Journal of Image and Graphics, 2019, 24(4):550-561.
15 徐刚, 岳继光, 董延超, 娄启佳, 等. 深度卷积网络卫星图像水泥厂目标检测. 中国图象图形学报,2019;24(4):550-561.
16 Cai Bowen, Wang Shugen, Wang Lei, et al. Extraction of Urban Impervious Surface from High-resolution Remote Sensing Imagery based on Deep Learning [J]. Journal of Geo-information Science,2019,21(9):1420-1429.
16 蔡博文,王树根,王磊,等.基于深度学习模型的城市高分辨率遥感影像不透水面提取[J]. 地球信息科学学报,2019,21(9):1420-1429.
17 Chen W T, Li X J, Wang L Z. Fine Land Cover Classification in an Open Pit Mining Area Using Optimized Support Vector Machine and WorldView-3 Imagery[J]. Remote Sensing, 2020, 12:82. .
18 Peng Y, Zhang Z M, He G J, et al. An Improved GrabCut Method based on a Visual Attention Model for Rare-earth Ore Mining Area Recognition with High-resolution Remote Sensing Images[J]. Remote Sensing, 2019, 11:987. doi:10.3390/rs11080987.
doi: 10.3390/rs11080987
19 Zhu Guixiang. Ecological Compensation:the Need for Mineral Resources Exploitation and Environmental Restoration of Henan Province[J]. Natural Resource Economics of China, 2010,23(5):25-27.
19 朱桂香. 生态补偿:河南矿产资源开发与环境修复的需求[J]. 中国国土资源经济,2010,23(5):25-27.
20 Benz U C, Hofmann P, Willhauck G, et al. Multi-resolution, Object-oriented Fuzzy Analysis of Remote Sensing Data for GIS-ready Information[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2004,58(3-4):239-258.
21 Huang Xin. Multiscale Texture and Shape Feature Extraction and Object-oriented Classification for very High Resolution Remotely Sensed Imagery[D]. Wuhan: Wuhan university, 2009.
21 黄昕. 高分辨率遥感影像多尺度纹理、形状特征提取与面向对象分类研究[D]. 武汉:武汉大学, 2009.
22 Chen T, Trinder J C, Niu R Q. Object-oriented Landslide Mapping Using ZY-3 Satellite Imagery, Random Forest and Mathematical Morphology, for the Three-Gorges Reservoir, China[J]. Remote Sensing, 2017,9(4):333. doi:10.3390/rs9040333.
doi: 10.3390/rs9040333
23 LeCun Y, Kavukcuoglu K, Farabet C, Convolutional Networks and Applications in Vision[C]∥ IEEE International Symposium on Circuits and Systems, 2010: 253-256.
24 Srivastava N, Hinton G, Krizhevsky A, et al. Dropout: A Simple Way to Prevent Neural Networks from Overfitting[J]. Journal of Machine Learning Research, 2014,15:1929-1958.
25 Cuides Developer. Keras: The Python Deep Learning Library[EB/OL]. ,2021.
26 Cortes C, Vapnik V. Support-vector Networks[J]. Machine learning, 1995,20(3):273-297.
27 Wu Jianping, Yang Xingwei. Accuracy Analysis of Classification of Remotely Sensed Data[J].Remote Sensing Technology and Application,1995,10(1):17-24.
27 吴健平,杨星卫.遥感数据分类结果的精度分析[J].遥感技术与应用,1995,10(1):17-24.
[1] 刘钊,赵桐,廖斐凡. 基于卷积神经网络预测结果的缝隙修复算法研究[J]. 遥感技术与应用, 2021, 36(2): 247-255.
[2] 陈敏,潘佳威,李江杰,徐璐,刘加敏,韩健,陈奕云. 结合VGGNet与Mask R-CNN的高分辨率遥感影像建设用地检测[J]. 遥感技术与应用, 2021, 36(2): 256-264.
[3] 徐瑾昊,冯敏,王建邦,冉有华,祁元,杨联安,李新. 基于高分遥感数据和深度学习的石冰川自动提取研究[J]. 遥感技术与应用, 2020, 35(6): 1329-1336.
[4] 张永宏,曹海啸,阚希. 基于FY-4A/AGRI时空特征融合的新疆地区积雪判识[J]. 遥感技术与应用, 2020, 35(6): 1337-1347.
[5] 郑琪,邸苏闯,潘兴瑶,刘洪禄,朱永华,张岑,周星. 基于Rapid Eye数据的北京生态涵养区土地利用分类及变化研究[J]. 遥感技术与应用, 2020, 35(5): 1118-1126.
[6] 牟昱璇,邬明权,牛铮,黄文江,杨尽. 南方地区复杂条件下的耕地面积遥感提取方法[J]. 遥感技术与应用, 2020, 35(5): 1127-1135.
[7] 杨瑞,祁元,苏阳. 深度学习U-Net方法及其在高分辨卫星影像分类中的应用[J]. 遥感技术与应用, 2020, 35(4): 767-774.
[8] 连喜红,祁元,王宏伟,张金龙,杨瑞. 基于面向对象的青海湖环湖区居民地信息自动化提取[J]. 遥感技术与应用, 2020, 35(4): 775-785.
[9] 郝睿,李兆富,张舒昱,潘剑君,姜小三,张文敏,宋金超. 整合无人机和面向对象的农村居住环境信息提取[J]. 遥感技术与应用, 2020, 35(3): 576-586.
[10] 王芳,杨武年,王建,谢兵,任金铜. 遥感影像多尺度分割中最优尺度的选取及评价[J]. 遥感技术与应用, 2020, 35(3): 623-633.
[11] 张峰极,吴艳兰,姚雪东,梁泽毓. 基于改进DenseNet网络的多源遥感影像露天开采区智能提取方法[J]. 遥感技术与应用, 2020, 35(3): 673-684.
[12] 郑宗生,胡晨雨,黄冬梅,邹国良,刘兆荣,宋巍. 基于迁移学习及气象卫星云图的台风等级分类研究[J]. 遥感技术与应用, 2020, 35(1): 202-210.
[13] 李强,冯德俊,瑚敏君,伍燚垚,杨历辉. 集成特征分量的高分二号影像阴影检测[J]. 遥感技术与应用, 2019, 34(6): 1252-1260.
[14] 李越帅,郑宏伟,罗格平,杨辽,王伟胜,桂东伟. 集成U-Net方法的无人机影像胡杨树冠提取和计数[J]. 遥感技术与应用, 2019, 34(5): 939-949.
[15] 刘天福,陈学泓,董琪,曹鑫,陈晋. 深度学习在GlobeLand30-2010产品分类精度优化中应用研究[J]. 遥感技术与应用, 2019, 34(4): 685-693.