城市建成区多源遥感协同提取方法研究

发布时间：2018-04-19 00:19

本文选题：DMSP/OLS夜光数据 + 城市建成区　；参考：《地球信息科学学报》2017年11期

【摘要】：城市建成区边界是城市研究重要的基础信息,也是落实城市功能空间布局、实施界限管控的前提。DMSP/OLS夜间灯光数据已被广泛应用于城市建成区的提取,但由于受饱和、扩散及低分辨率问题的影响,导致仅依靠DMSP/OLS NTL映射城市建成区仍然是一个巨大的挑战。本文以京津冀为例,采用MODIS NDVI和DMSP/OLS夜光数据相结合解决NTL影像的饱和及扩散问题,提取城市建成区潜在范围,并辅以Landsat NDVI数据,采用本文提出的最大自相关双阈值方法进行自适应修正,最后采用目视解译对结果进行验证。实验结果表明,多源遥感协同方法提取城市建成区的总体精度和kappa系数分别为92.9%和0.88,在空间分布和统计数据中均有较高的有效性和可靠性。
[Abstract]:The borderline of urban built-up area is the important basic information of urban research, and it is also the premise of carrying out the layout of urban functional space, and the premise of carrying out boundary control. DMSP / OLS night light data have been widely used in extracting urban built-up area, but because of saturation,Because of the influence of diffusion and low resolution problems, it is still a great challenge to rely only on DMSP/OLS NTL to map urban areas.In this paper, taking Beijing-Tianjin-Hebei as an example, the problem of saturation and diffusion of NTL images is solved by combining MODIS NDVI and DMSP/OLS nocturnal light data, and the potential range of urban built-up areas is extracted, supplemented by Landsat NDVI data.The maximum autocorrelation double threshold method proposed in this paper is used for adaptive correction and visual interpretation is used to verify the results.The experimental results show that the total precision and kappa coefficient of multi-source remote sensing cooperative method are 92.9% and 0.88 respectively, which are effective and reliable in spatial distribution and statistical data.
【作者单位】：中国科学院新疆生态与地理研究所荒漠与绿洲生态国家重点实验室;中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室;中国科学院大学;
【基金】：国家重点研发计划项目(2016YFB0501404) 国家自然科学基金项目(41671436、41421001)
【分类号】：P237

【相似文献】