Extraction of land use information in various karst landscapes based on multiple scale-spectral differential subdivision of GF-1 images
-
摘要: 影像分割是高分辨率遥感影像信息提取的前提,遥感影像分割的精确程度直接影响遥感分类的精度。为提高喀斯特山区遥感影像信息提取的精度,采用多尺度-光谱差异分割对喀斯特山区高分辨率影像分割,通过标准最近邻分类法提取土地利用信息,对比了仅多尺度分割、多尺度-光谱差异分割两种分割方法下喀斯特山区土地利用信息提取的精度。结果表明:(1)多尺度-光谱差异分割能改善过分割和欠分割现象。(2)多尺度-光谱差异分割优于单一使用多尺度分割。(3)多尺度-光谱差异分割综合了影像的光谱、纹理、形状等特征,进而提高了喀斯特山区影像分割分类的精度。
-
关键词:
- GF-1影像 /
- 喀斯特 /
- 多尺度-光谱差异分割 /
- 土地利用信息
Abstract: Image segmentation is a necessary step in information extraction from high-resolution images. The accuracy of such division can directly influence the precision of remote sensing classification. This work uses multiple scale-spectral differential method to conduct the division in karst mountainous areas, thus enhances the accuracy of information extraction. Using the standard most-adjacent classification method, information of land use is extracted from divided images. The accuracy of land use information extraction is compared for only multiple-scale subdivision and multiple-scale spectral differential subdivision. Results demonstrate that (1) multiple scale-spectral difference subdivision can solve the problems of over-division and under-division. (2) Multiple scale-spectral difference subdivision is superior to only using multiple scale subdivision. (3) Multiple scale-spectral difference subdivision considers many features of images such as spectra, lamination, and shape, thus permits to enhance the accuracy of division and classification of images in karst mountainous areas. -
[1] Lv Zhiyong, Zhang Penglin, Jón Atli Benediktsson. Automatic Object-Oriented, Spectral-Spatial Feature Extraction Driven by Tobler’s First Law of Geography for Very High Resolution Aerial Imagery Classification[J].Remote Sensing.2017,9(3):1-17. [2] 吴文欢,于宏,赵英俊,等.基于面向对象的高分辨率遥感影像目标信息提取[J].世界核地质科学,2016,33(2):91-95. [3] 刘小鹏,窦关新,赵宝军,等.面向对象的遥感影像信息提取研究[J].北京测绘,2016(2):106-109. [4] 佃袁勇,方圣辉,姚崇怀.多尺度分割的高分辨率遥感影像变化检测[J].遥感学报,2016,20(1):129-137. [5] 王文泉,陈永富,李肇晨,等.基于面向对象的热带林分类方法研究[J].南京林业大学学报(自然科学版),2017,41(3):117-123. [6] 龚文峰,刘芳平,王笑峰,等.基于面向对象的太阳岛土地利用信息提取研究[J].黑龙江水利,2016,2(10):28-34. [7] 任传帅,叶回春,崔贝,等.基于面向对象分类的芒果林遥感提取方法研究[J].资源科学,2017,39(8):1584-1591. [8] Sun Wenyi, Tian Yuansheng, Mu Xingmin, et al. Loess Landslide Inventory Map Based on GF-1 Satellite Imagery[J]. Remote Sensing 2017,9(4):314-330. [9] 黄慧萍,吴炳方,李苗苗,等.高分辨率影像城市绿地快速提取技术与应用[J].遥感学报,2004(1):68-74. [10] 周亦,张亚亚.利用eCognition进行高分一号卫星数据土地利用现状解译能力测试[J].测绘通报,2016(8):77-80+94. [11] Chen Jie, Deng Min, Xiao Pengfeng, et al. Optimal spatial scale choosing for high resolution imagery based on texture frequency analysis[J].Journal of Remote Sensing,2011,15(3):492-511. [12] 费鲜芸,王婷,魏雪丽.基于多尺度分割的遥感影像滨海湿地分类[J].遥感技术与应用,2015,30(2):298-303. [13] 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(4):521-530. [14] 林雨准,张保明,徐俊峰,等.多特征多尺度相结合的高分辨率遥感影像建筑物提取[J].测绘通报,2017(12):53-57. [15] 李雪冬,张洪岩,杨广斌,等.辅以地貌类型的喀斯特地区土地利用信息提取[J].国土资源遥感,2016,28(3):138-145. [16] 张正健,李爱农,雷光斌,等.基于多尺度分割和决策树算法的山区遥感影像变化检测方法:以四川攀西地区为例[J].生态学报,2014,34(24):7222-7232. [17] Feng Xia, Qiu Kun, Cui Wenhong,et al. Multiscale description and recognition of target shape in high-resolution remote sensing images[J].Journal of Remote Sensing,2014,18(1):90-104. [18] 马燕妮,明冬萍,杨海平.面向对象影像多尺度分割最大异质性参数估计[J].遥感学报.2017,21(4):566-578. [19] 田甜,范文义,卢伟,等.面向对象的优势树种类型信息提取技术[J].应用生态学报,2015,26(6):1665-1672. [20] 刘佳雨.利用面向对象的信息提取技术进行城市用地分类[J].西部资源,2016(4):156-159+171. [21] 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. [22] 刘家福,刘吉平,姜海玲.eCognition 数字图像处理方法[M].北京:科学出版社,2017:52-56. [23] 安霞霞,杨广斌,许元红,等.喀斯特山区居民地多尺度遥感信息提取精度对比分析[J].中国岩溶,2017,36(4):501-511. [24] Zhai Yongguang, Qu Zhongyi, and Hao Lei. Land Cover Classification Using Integrated Spectral,Temporal,and Spatial Features Derived from Remotely Sensed Images[J].Remote Sensing,2018,10(3):383-408.
点击查看大图
计量
- 文章访问数: 1790
- HTML浏览量: 532
- PDF下载量: 429
- 被引次数: 0