地球信息科学学报 ›› 2016, Vol. 18 ›› Issue (3): 353-361.doi: 10.3724/SP.J.1047.2016.00353

• 扶贫空间信息系统关键技术研究专栏 • 上一篇    下一篇

高分辨率遥感土地覆盖分类技术的应用研究——以重庆市黔江贫困区为例

吴田军1,3(), 骆剑承2, 沈占锋2, 夏列钢4, 徐楠2   

  1. 1. 长安大学理学院数学与信息科学系,西安 710064
    2. 中国科学院遥感与数字地球研究所,北京 100101
    3. 浙江省海洋大数据挖掘与应用重点实验室,舟山 316022
    4. 浙江工业大学计算机学院,杭州 310023
  • 收稿日期:2015-09-07 修回日期:2015-10-20 出版日期:2016-03-10 发布日期:2016-03-17
  • 作者简介:

    作者简介:吴田军(1986-),男,博士,研究方向为遥感信息计算.E-mail:wutianjun1986@163.com

  • 基金资助:
    国家科技支撑计划项目"扶贫空间信息系统关键技术及其应用"第一课题"贫困地区资源环境监测评估与生态价值评价技术"的专题一和专题五(2012BAH33B01-01,2012BAH33B01-05);浙江省海洋大数据挖掘与应用重点实验室开放课题项目(OBDMA201508)

Application of Land Cover Classification Techniques in Poverty-stricken Areas Using High Resolution Remote Sensing: A Case Study of Qianjiang District of Chongqing City

WU Tianjun1,3,*(), LUO Jiancheng2, SHEN Zhanfeng2, XIA Liegang4, XU Nan2   

  1. 1. Department of mathematics and Information Science, College of Science, Chang'an University, Xi'an 710064, China
    2. Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100101, China
    3. Key Laboratory of Oceanographic Big Data Mining & Application of Zhejiang Province, Zhoushan 316022, China
    4. College of Computer Science and Technology, Zhejiang University of Technology, Hangzhou 310026, China
  • Received:2015-09-07 Revised:2015-10-20 Online:2016-03-10 Published:2016-03-17
  • Contact: WU Tianjun E-mail:wutianjun1986@163.com

摘要:

针对贫困区生态环境与资源的地表覆盖精细化调查需求,本文利用高分辨率遥感影像开展了土地覆盖信息提取的方法和应用研究.重点分析了高分辨率影像均值漂移分割,多特征提取与分析,对象级样本采集以及监督分类等技术,并综合实现了流程化的对象级土地覆盖分类.结果表明,本文串联的高分辨率影像分类技术能生成较精细的土地覆盖专题图,可及时为贫困区生态资源环境评价,碳核算等应用提供较可靠的地表覆盖数据.

关键词: 贫困区, 精细化土地覆盖制图, 高分辨率遥感影像, 面向对象分类

Abstract:

To meet the actual needs of refined land-cover information in poverty-stricken areas' ecological environment assessment and resource survey, this paper studies the technology and application of accurate land-cover classification based on the high-resolution satellite images. This article focuses on the technologies of high-resolution image segmentation using the mean shift algorithm, multiple feature extraction and analysis, automatic object-sample selection, and supervised classification. Then, we give a complete realization of the streamlined object-based land-cover classification by lining up the above technologies. Finally, taking Qianjiang district of Chongqing city as an example, we carry out a land cover information extraction experiment in its poverty-stricken areas. The experimental results show that the accurate and detailed land-cover maps can be obtained using the provided technology. The meticulous and reliable land-cover information can promptly provide a base-data-support for the eco-environmental assessment and carbon accounting applications in the poverty-stricken areas.

Key words: poverty-stricken areas, refined land-cover mapping, high-resolution remote sensing images, object-based classification