面向对象的喀斯特地区石漠化遥感信息提取研究——以贵州省大方地区为例

周迪; 倪忠云; 杨振宇

doi:10.12090/j.issn.1006-6616.2018.24.02.028

面向对象的喀斯特地区石漠化遥感信息提取研究——以贵州省大方地区为例

doi: 10.12090/j.issn.1006-6616.2018.24.02.028

周迪^1,,
倪忠云^{2, 3, 4, ,},
杨振宇¹

1.
首都师范大学地球空间信息科学与技术国际化示范学院, 北京 100048
2.
首都师范大学资源环境与旅游学院, 北京 100048
3.
成都理工大学旅游与城乡规划学院, 四川成都 610059
4.
地质灾害防治与地质环境保护国家重点实验室, 四川成都 610059

基金项目:

四川省教育厅项目 16ZA0080

2016年成都理工大学校级教学改革及课程建设项目 11100-16z056919

详细信息

作者简介:
周迪(1994-), 女, 在读硕士, 研究方向:InSAR技术监测地面沉降。E-mail:zhoudi_cnu@163.com

通讯作者:
倪忠云(1982-), 女, 博士, 讲师, 从事生态环境遥感和遥感地质等研究。E-mail:theoneyun@gmail.com

中图分类号: TP79;P391.5
计量
- 文章访问数: 50
- HTML全文浏览量: 12
- PDF下载量: 8
- 被引次数: 0
出版历程
- 收稿日期: 2017-09-29
- 修回日期: 2018-02-16
- 刊出日期: 2018-04-28

OBJECT-ORIENTED REMOTE SENSING INFORMATION EXTRACTION METHOD FOR ROCKY DESERTIFICATION IN KARST AREA-A CASE STUDY OF DAFANG COUNTY, GUIZHOU

ZHOU Di^1
,,
NI Zhongyun^{2, 3, 4
, ,},
YANG Zhenyu¹

1.
College of Geospatial information Science and Technology, Capital Normal University, Beijing 100048, China
2.
College of Resource Environment and Tourism, Capital Normal University, Beijing 100048, China
3.
College of Tourism and Urban-Rural Planning, Chengdu University of Technology, Chengdu 610059, Sichuan, China
4.
State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of Technology, Chengdu 610059, Sichuan, China

摘要

摘要: 针对现有基于像素的监督和非监督分类方法在地质环境复杂、地形起伏较大、阴影明显的喀斯特石漠化地区难以满足石漠化信息提取精度要求的问题，采用基于纹理特征数据和地形数据辅助面向对象方法进行喀斯特地区石漠化信息的提取。该方法首先依据石漠化分布在TM/ETM+影像面积大小不均匀的特征，利用纹理和地形因子计算最优分割参数进行多尺度分割；然后根据植被覆盖率、岩石裸露率以及坡度因子构建石漠化分级指标；最后参照石漠化分级标准、光谱信息以及纹理特征等建立的分类规则提取喀斯特地区石漠化信息。选取贵州省石漠化严重的大方县时序TM/ETM+影像进行石漠化信息提取试验，结果表明：与基于像素的监督分类和非监督分类方法相比，基于面向对象的分类可以有效地减少因复杂地形导致石漠化信息提取结果"椒盐化"现象，提取精度明显优于基于像素的监督分类和非监督分类方法。
- 喀斯特石漠化 /
- 面向对象分类 /
- 监督分类 /
- 非监督分类 /
- 信息提取 /
- 大方地区
Abstract: The existing pixel-based supervised and unsupervised classification methods can't meet the requirements of rocky desertification information extraction accuracy in karst rocky desertification area under the circumstances of complicated geological environment, large topographic relief and obvious shadows. In order to improve the accuracy of remote sensing image information extraction, texture feature data and topographic data are used to assist the object oriented method in the rocky desertification information extraction in karst rocky desertification area. Firstly, based on the characteristics of rocky desertification with uneven image sizes in TM/ETM+, the optimal segmentation parameters are calculated using texture and terrain factors to conduct multi-scale segmentation. Secondly, the grading indexes of rocky desertification are established based on vegetation coverage rates, rock exposure rates and slope factors. Finally, according to the grading rules of rocky desertification, spectral information and texture features, the information of rocky desertification in Karst area is extracted. The temporal TM/ETM+images of rocky desertification areas in DaFang, Guizhou, are selected for rocky desertification information extraction. The results show that comparing with pixel-based supervised classification and unsupervised classification methods, the object-oriented classification technology can effectively reduce the "salt and pepper phenomenon" caused by complicated topography, and the extraction accuracy is much better.
- karst rocky desertification /
- object-oriented classification /
- supervised classification /
- unsupervised classification /
- information extraction /
- Dafang County

HTML全文

图 1 研究区位置图

Figure 1. Location map of the study area

下载: 全尺寸图片幻灯片

图 2 石漠化遥感影像信息提取流程图

Figure 2. Flow chart of rocky desertification remote sensing image information extraction

下载: 全尺寸图片幻灯片

图 3 纹理特征数据

Figure 3. Texture feature data

下载: 全尺寸图片幻灯片

图 4 坡度和NDVI信息提取结果

Figure 4. Results of gradient and NDVI information extraction

下载: 全尺寸图片幻灯片

图 5 均值方差曲线

Figure 5. Mean variance curve

下载: 全尺寸图片幻灯片

图 6 分割结果对比图

a—纹理特征数据和地形数据未参与分割；b—纹理特征数据和地形数据参与分割

Figure 6. Comparison charts of segmentation results

下载: 全尺寸图片幻灯片

图 7 研究区石漠化信息提取结果对比图

Figure 7. Comparison of information extraction results of rocky desertification in study area

下载: 全尺寸图片幻灯片

图 8 研究区石漠化分布图

Figure 8. Distribution of rocky desertification in study area

下载: 全尺寸图片幻灯片

图 9 实地验证区域图

Figure 9. Maps of field verification

下载: 全尺寸图片幻灯片

图 10 实地验证石漠化区域照片

Figure 10. Pictures of field verification in desertification areas

下载: 全尺寸图片幻灯片

表 1 分割参数的设置

Table 1. Settings of segmentation parameters

分割层次	波段选择	分割尺度	异质性因子
分割层次	波段选择	分割尺度	光谱权重	形状权重	光滑度	紧密度
Level1	741+4纹理	20	0.7	0.3	0.5	0.4
Level2	741+4纹理	50	0.7	0.3	0.5	0.4
Level3	741+4纹理	100	0.7	0.3	0.5	0.4

下载: 导出CSV

表 2 石漠化等级划分标准^[30]

Table 2. Grading standards of rock desertification ^[30]

强度等级	坡度/%	岩石裸露率/%	植被覆盖率/%	利用价值
无石漠化	< 15	< 10	>70	宜农宜林牧
微度石漠化	>15	10~30	50~70	宜林牧
轻度石漠化	>18	30~50	35~50	临界宜林牧
中度石漠化	>22	50~70	10~30	难利用地
重度石漠化	>25	>70	< 10	无利用价值

下载: 导出CSV

表 3 面向对象的喀斯特地区石漠化遥感信息提取规则

Table 3. Rules of object-oriented remote sensing information extraction method for rocky desertification in Karst area

分割层次	分割尺度	提取信息	规则(模糊分类定义的成员函数)
Level1	20	水体喀斯特地区非喀斯特地区	200≤Ratio(B4)≤800, Brightness < 180, length/width > 3.28, Mean(con) < 0.34; Ratio(B4)≥800, Ratio(B2)≤1000, NDVI≥0.405, Shape Index > 0.62; NDVI < 0.405, Ratio(B2)>1000, Mean(con) < 5;
Level2	50	无石漠化有石漠化	NDVI≥0.525, Brightness≤1300, length/width > 1.5, Mean(slope) < 15; NDVI < 0.525, Brightness>1300, Stdv(B2)+Stdv(B3)+Stdv(B4) < 20, 15 < Mean(slope) < 45;
Level3	100	中度石漠化轻度石漠化重度石漠化微度石漠化	Ratio(B2) < 4.45, Stdv(B4) < 12, NDVI < -0.25, Mean(slope)>22; 200 < Brightness < 400, Mean(slope)>18, 1.5 < Mean(Ent) < 2.6; Ratio(B3) < 2.26, Mean(slope)>25, Mean(Ent)>2.6, Mean(con)>5; 0.25 < Ratio(B7) < 0.31, Mean(slope)>15, Mean(con)>7;
注：Brightness为对象平均的多光谱灰度值，length/width为对象的长宽比，Stdv标准差，Ratio为比率，Mean(con)为对比度均值，Mean(Ent)为熵均值，Mean(slope)为坡度均值，NDVI=(Nir-red)/(Nir+red)，Shape Index=e/4√A(e为边长，A为面积)，B2、B3、B4、B7是TM/ETM+多光谱影像的波段。

下载: 导出CSV

表 4 1988—2016年大方地区喀斯特石漠化面积

Table 4. Area of karst rocky desertification in Dafang from 1988 to 2016

	1988年		2002年		2016年
	面积/km²	比例/%	面积/km²	比例/%	面积/km²	比例/%
微度石漠化区	139.27	5.07	89.67	3.27	77.64	2.83
轻度石漠化区	78.5	2.86	135.8	4.95	144.43	5.26
中度石漠化区	74.15	2.70	66.35	2.42	61.91	2.25
重度石漠化区	37.28	1.36	30.6	1.11	19.78	0.72
水系	73.81	2.69	75.39	2.75	74.24	2.70
无石漠化区	1746.29	63.59	1751.49	63.78	1771.3	64.50
非喀斯特地层区	596.7	21.73	596.7	21.73	596.7	21.73
总计	2746	100.00	2746	100.00	2746	100.00

下载: 导出CSV

表 5 分类方法精度对比表

Table 5. Comparison of the accuracy of each method

方法	监督分类			非监督分类			面向对象分类
年份	1988年	2002年	2016年	1988年	2002年	2016年	1988年	2002年	2016年
总体精度	72.15%	78.81%	76.43%	62.50%	63.25%	72.90%	91.45%	93.56%	95.31%
Kappa系数	0.65	0.71	0.68	0.54	0.55	0.58	0.9	0.92	0.94

下载: 导出CSV

参考文献(30)

[1]	周忠发.遥感和GIS技术在贵州喀斯特地区土地石漠化研究中的应用[J].水土保持通报, 2001, 21(3):52~54. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=stbctb200103014 ZHOU Zhongfa. Application of remote sensing and GIS technology for land desertification in Guizhou Karst Region[J]. Bulletin of Soil and Water Conservation, 2001, 21(3):52~54. (in Chinese with English abstract) http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=stbctb200103014
[2]	王金华. 近30年来粤北土地石漠化动态变化及发展趋势研究[D]. 广州: 华南师范大学, 2007. http://cdmd.cnki.com.cn/Article/CDMD-10574-2007184881.htm WANG Jinhua. Study on dynamic change and development trend of rocky desertification in Northern Guangdong in recent 30 years[D]. Guangzhou: South China Normal University, 2007. (in Chinese) http://cdmd.cnki.com.cn/Article/CDMD-10574-2007184881.htm
[3]	陈起伟, 熊康宁, 兰安军.基于3S的贵州喀斯特石漠化遥感监测研究[J].干旱区资源与环境, 2014, 28(3):62~67. http://kns.cnki.net/KCMS/detail/detail.aspx?filename=ghzh201403011&dbname=CJFD&dbcode=CJFQ CHEN Qiwei, XIONG Kangning, LAN Anjun. Monitoring studies on Karst rocky desertification in Guizhou based on 3S[J]. Journal of Arid Land Resources and Environment, 2014, 28(3):62~67. (in Chinese with English abstract) http://kns.cnki.net/KCMS/detail/detail.aspx?filename=ghzh201403011&dbname=CJFD&dbcode=CJFQ
[4]	Gill T K, Phinn S R, Armston J D, et al. Estimating tree-cover change in Australia:Challenges of using the MODIS vegetation index product[J]. International Journal of Remote Sensing, 2009, 30(6):1547~1565. doi: 10.1080/01431160802509066
[5]	岳跃民, 张兵, 王克林, 等.石漠化遥感评价因子提取研究[J].遥感学报, 2011, 15(4):722~736. http://www.wanfangdata.com.cn/details/detail.do?_type=degree&id=Y1826234 YUE Yuemin, ZHANG Bing, WANG Kelin, et al. Remote sensing of indicators for evaluating karst rocky desertification[J]. Journal of Remote Sensing, 2011, 15(4):722~736. (in Chinese with English abstract) http://www.wanfangdata.com.cn/details/detail.do?_type=degree&id=Y1826234
[6]	Townsend P A, Walsh S J. Remote sensing of forested wetlands:Application of multitemporal and multispectral satellite imagery to determine plant community composition and structure in southeastern USA[J]. Plant Ecology, 2001, 157(2):129~149. https://uncch.pure.elsevier.com/en/publications/remote-sensing-of-forested-wetlands-application-of-multitemporal-
[7]	李朝阳, 况顺达, 李志忠, 等.喀斯特石漠化遥感信息增强与监测技术[J].桂林工学院学报, 2007, 27(1):30~35. http://d.old.wanfangdata.com.cn/Periodical/glgxy200701006 LI Zhaoyang, KUANG Shunda, LI Zhizhong, et al. RS monitoring technology in karst rocky desertification[J]. Journal of Guilin University of Technology, 2007, 27(1):30~35. (in Chinese with English abstract) http://d.old.wanfangdata.com.cn/Periodical/glgxy200701006
[8]	杨奇勇, 蒋忠诚, 马祖陆, 等.基于地统计学和遥感的岩溶区石漠化空间变异特征[J].农业工程学报, 2012, 28(4):243~247. doi: 10.3969/j.issn.1002-6819.2012.04.040 YANG Qiyong, JIANG Zhongcheng, MA Zulu, et al. Spatial variability of karst rock desertification based on geostatistics and remote sensing[J]. Transactions of the CSAE, 2012, 28(4):243~247. (in Chinese with English abstract) doi: 10.3969/j.issn.1002-6819.2012.04.040
[9]	Gambarova E, Gambarov A, Ismayilov J. Applying neural networks in rare vegetation communities classification of remotely sensed images[J]. Optical Memory and Neural Networks, 2008, 17(2):157~166. doi: 10.3103/S1060992X08020100
[10]	胡宝清, 王世杰.基于3S技术的区域喀斯特石漠化过程、机制及风险评估:以广西都安为例[M].北京:科学出版社, 2008. HU Baoqing, WANG Shijie. Process, Mechanism and risk assessment of regional Karst desertification based on 3S technology:a case study of Du'an, Guangxi[M]. Beijing:Science Press, 2008. (in Chinese)
[11]	Willhauck G, Schneider T, De Kok R, et al. Comparison of object oriented classification techniques and standard image analysis for the use of change detection between SPOT multispectral satellite images and aerial photos[A]. Proceedings of XIX ISPRS Congress[C]. Amsterdam: IAPRS, 2000.
[12]	贾明明, 任春颖, 刘殿伟, 等.基于环境星与MODIS时序数据的面向对象森林植被分类[J].生态学报, 2014, 34(24):7167~7174. http://www.ecologica.cn/stxb/ch/html/2014/24/stxb201310112438.htm JIA Mingming, REN Chunying, LIU Dianwei, et al. Object-oriented forest classification based on combination of HJ-1 CCD and MODIS-NDVI data[J]. Acta Ecologica Sinica, 2014, 34(24):7167~7174. (in Chinese with English abstract) http://www.ecologica.cn/stxb/ch/html/2014/24/stxb201310112438.htm
[13]	谢雨萍, 吴虹, 刘泽东, 等.恭城县岩溶石漠化环境变化定量遥感研究[J].桂林工学院学报, 2009, 29(1):65~71. http://d.old.wanfangdata.com.cn/Periodical/ghqzyyhj201403011 XIE Yuping, WU Hong, LIU Zedong, et al. Environment change of karst rocky desertification by quantitative remote sensing in Gongcheng[J]. Journal of Guilin University of Technology, 2009, 29(1):65~71. (in Chinese with English abstract) http://d.old.wanfangdata.com.cn/Periodical/ghqzyyhj201403011
[14]	Lupo F, Linderman M, Vanacker V, et al. Categorization of land-cover change processes based on phenological indicators extracted from time series of vegetation index data[J]. International Journal of Remote Sensing, 2007, 28(11):2469~2483. doi: 10.1080/01431160600921943
[15]	孟小军, 莫源富.基于遥感技术的打狗河流域1999-2009年间石漠化及绿化研究[J].桂林理工大学学报, 2013, 33(4):622~628. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=glgxy201304008 MENG Xiaojun, MO Yuanfu. Rocky desertification and afforestation of Dagou River basin during 1999-2009 based on remote sensing technology[J]. Journal of Guilin University of Technology, 2013, 33(4):622~628. (in Chinese with English abstract) http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=glgxy201304008
[16]	童立强.西南岩溶石山地区石漠化信息自动提取技术研究[J].国土资源遥感, 2003, (4):35~38. doi: 10.6046/gtzyyg.2003.04.09 TONG Liqiang. A method for extracting remote sensing information from rocky desertification areas in Southwest China[J]. Remote Sensing for Land & Resources, 2003, (4):35~38. (in Chinese with English abstract) doi: 10.6046/gtzyyg.2003.04.09
[17]	林雪. 面向林地信息的高分一号遥感影像融合与分类研究[D]. 北京: 北京林业大学, 2016. http://cdmd.cnki.com.cn/Article/CDMD-10022-1016145061.htm LIN Xue. Study on fusion algorithms and classincation methods for GF-1 data oriented to forestland information[D]. Beijing: Beijing Forestry University, 2016. (in Chinese with English abstract) http://cdmd.cnki.com.cn/Article/CDMD-10022-1016145061.htm
[18]	YANG Panpan. Research on object-oriented vegetation classification method based on texture features of high resolution remote sensing images[J]. Kunming:Yunnan Normal University, 2017. (in Chinese with English abstract)
[19]	梁茂昆. 基于面向对象及Landsat影像的广州市城市空间格局演变研究[D]. 抚州: 东华理工大学, 2017. LIANG Maokun. Research on the evolution of Guangzhou urban spatial pattern based on object-oriented and Landsat images[D]. Fuzhou: East China University of Technology, 2017. (in Chinese with English abstract)
[20]	李海霞. 高分辨率遥感影像对象分类方法研究及其城乡规划监测应用[D]. 北京: 中国农业大学, 2014. http://cdmd.cnki.com.cn/Article/CDMD-10019-1014223423.htm LI Haixia. Study on the methods of object classification and its application on urban and rural plan monitoring with high spatial resolution remotely sensed data[D]. Beijing: China Agricultural University, 2014. (in Chinese with English abstract) http://cdmd.cnki.com.cn/Article/CDMD-10019-1014223423.htm
[21]	陈建容. 面向对象的土地利用/覆被信息提取——以乐山市GF-1号影像为例[D]. 成都: 成都理工大学, 2016. http://www.wanfangdata.com.cn/details/detail.do?_type=degree&id=Y3048461 CHEN Jianrong. Object-oriented land use/cover information extraction[D]. Chengdu: Chengdu University of Technology, 2016. (in Chinese) http://www.wanfangdata.com.cn/details/detail.do?_type=degree&id=Y3048461
[22]	贾伟. 面向对象的复杂地形区土地利用信息提取研究[D]. 西宁: 青海师范大学, 2015. http://cdmd.cnki.com.cn/Article/CDMD-10746-1015661361.htm JIA Wei. Land use information extraction for complicated terrain regions from object-oriented classification technique[D]. Xining: Qinghai Normal University, 2015. (in Chinese with English abstract) http://cdmd.cnki.com.cn/Article/CDMD-10746-1015661361.htm
[23]	欧阳华璘, 沈敬伟, 周廷刚.面向对象分类方法在台风灾害信息提取中的应用研究[J].自然灾害学报, 2016, 25(6):9~17. http://kns.cnki.net/KCMS/detail/detail.aspx?filename=zrzh201606002&dbname=CJFD&dbcode=CJFQ OUYANG Hualin, SHEN Jingwei, ZHOU Tinggang. Application of object-oriented classification method to typhoon disaster information extraction[J]. Journal of Natural Disaster, 2016, 25(6):9~17. (in Chinese with English abstract) http://kns.cnki.net/KCMS/detail/detail.aspx?filename=zrzh201606002&dbname=CJFD&dbcode=CJFQ
[24]	李雪冬, 杨广斌, 李蔓, 等.面向对象的喀斯特地区土地利用遥感分类信息提取-以贵州毕节地区为例[J].中国岩溶, 2013, 32(2):231~237. http://www.cqvip.com/QK/95841X/201302/46531896.html LI Xuedong, YANG Guangbin, LI Man, et al. RS classification information extraction of landuse in karst area by means of object oriented approach:a case in Bijie, Guizhou[J]. Carsologica Sinica, 2013, 32(2):231~237. (in Chinese with English abstract) http://www.cqvip.com/QK/95841X/201302/46531896.html
[25]	刘海龙. 面向对象的石漠化遥感监测及过程模拟研究[D]. 昆明: 昆明理工大学, 2015. http://cdmd.cnki.com.cn/Article/CDMD-10674-1015636041.htm LIU Hailong. Object-oriented remote sensing monitoring and process simulation of rocky desertification[D]. Kunming: Kunming University of Science and Technology, 2015. (in Chinese) http://cdmd.cnki.com.cn/Article/CDMD-10674-1015636041.htm
[26]	曹密媛. 基于遥感影像的地物要素智能识别与提取研究[D]. 西安: 长安大学, 2015. http://www.wanfangdata.com.cn/details/detail.do?_type=degree&id=D748355 CAO Miyuan. Based on the feature elements of intelligent recognition and extraction of remote sensing image[D]. Xi'an: Chang'an University, 2015. (in Chinese with English abstract) http://www.wanfangdata.com.cn/details/detail.do?_type=degree&id=D748355
[27]	佘宇晨. 基于GF-2光谱特征的石漠化信息自动提取[D]. 株洲: 中南林业科技大学, 2017. http://cdmd.cnki.com.cn/Article/CDMD-10538-1017117982.htm SHE Yuchen. Automatic extraction of desertification information based on GF-2 spectral characteristics[D]. Zhuzhou: Central South University of Forestry & Technology, 2017. (in Chinese with English abstract) http://cdmd.cnki.com.cn/Article/CDMD-10538-1017117982.htm
[28]	苏伟, 李京, 陈云浩, 等.基于多尺度影像分割的面向对象城市土地覆被分类研究-以马来西亚吉隆坡市城市中心区为例[J].遥感学报, 2007, 11(4):521~530. doi: 10.11834/jrs.20070472 SU Wei, LI Jing, CHEN Yunhao, et al. Object-oriented urban land-cover classification of multi-scale image segmentation method——a case study in Kuala Lumpur City center, Malaysia[J]. Journal of Remote Sensing, 2007, 11(4):521~530. (in Chinese with English abstract) doi: 10.11834/jrs.20070472
[29]	LIAO Ning, XU Lisha, QIAN Xiaoshan. Texture classification based on multi-scale wavelet[J]. Journal of System Simulation, 2015, 27(9):1951~1959. http://www.en.cnki.com.cn/Article_en/CJFDTotal-XTFZ201509007.htm
[30]	熊康宁.喀斯特石漠化的遥感-GIS典型研究[M].地质出版社, 2002 Xiong Kangning. A Typical Remote Sensing-GIS Study of Karst Rocky Desertification[M]. Geological Publishing House, 2002(in Chinese)