基于RS与GIS的南涧县土地覆被监测研究

发布时间：2018-10-15 12:57

【摘要】：随着全球环境变化的发展,我国对土地利用变化监测的需求与日俱增。以往采用中低分辨率国外卫星数据开展基于像素的土地覆被信息提取,分类精度不高;部分面向对象的土地利用信息提取局限于实验区开展研究,基于基本行政单元的土地利用遥感信息提取工程化应用不足。针对以往研究存在的局限性,本文主要采用国外卫星数据及国产卫星数据相结合的方式,选取2000年的ETM+和2014年GF-1号2期遥感数据为主要数据源,以南涧县县级行政区域为研究区,探索面向对象的土地覆被信息提取方法,并将其推广应用于县域土地覆被变化信息提取研究中。本文研究主要取得了以下成果:(1)本文以南涧县作为研究区,采用ETM+、GF-1号卫星遥感数据,借助ESP工具算法开展了面向对象的土地覆被最优分割尺度、形状因子和紧致度因子实验,最后确定南涧县面向对象技术方法指标体系为:ETM+影像最优分割尺度为20,GF-1号最优分割尺度为40,形状因子和紧致度因子参数均为0.4和0.6。(2)结合南涧县2000、2014年南涧县土地利用现状资料,对面向对象的土地覆被分类结果精度进行验证,其中,GF-1号分类结果总体精度为88.1293%,Kappa系数为0.8469;ETM+影像分类结果总体精度为77.0385%,Kappa系数为0.6925,更进一步印证了GF-1号土地覆被分类结果更高。据此,探索建立了县域实用性较强的GF-1号面向对象分类方法。(3)分析2000-2014年土地覆被动态变化信息,显示:南涧县2000年和2014年土地覆被类型均以林地为主,分别占整个研究区土地利用面积的46.36%和57.71%,其次是草地、耕地和建筑用地,未利用地和水域用地最少;林地的涨幅最大,为11.35%,未利用地的降幅最大,为11.95%,其他地类均有升降变化,但不明显;土地覆被程度综合指数从211增加到225.91,以每年1.065%的速度增加;土地覆被各类型均存在转入和转出的现象,未利用地转出的面积最大,林地最少,转入面积最大的是水域,最少的是耕地,转化面积都不大,整体来看相对平稳。在此基础上,分析了南涧县2000-2014年土地覆被变化存在林地、建筑用地及水域面积增加,耕地、草地面积减少,土地利用的水平不高,土地利用结构和布局不尽合理等问题。提出了统筹城乡发展、加强对水资源的利用效率、发展经济林产业,优先林业结构、加大保护耕地的力度、继续实施退耕还林还草政策等土地利用规划建议。
[Abstract]:With the development of global environmental change, the demand for land use change monitoring in China is increasing day by day. In the past, low-resolution foreign satellite data were used to extract land cover information based on pixels, and the classification accuracy was not high; some object-oriented land use information extraction was limited to the experimental area to carry out research. The engineering application of land use remote sensing information extraction based on basic administrative unit is insufficient. In view of the limitations of previous studies, this paper mainly adopts the combination of foreign satellite data and domestic satellite data, and selects the 2000 ETM and 2014 GF-1 as the main data sources. Taking the county administrative region of Nanjian county as the research area, this paper explores the method of extracting land cover information by object oriented method, and applies it to the research of extracting land cover change information in county area. The main achievements of this paper are as follows: (1) in this paper, Nanjian County is used as the research area, the ETM and GF-1 satellite remote sensing data are adopted, and the object-oriented experiments of optimal land cover segmentation scale, shape factor and compactness factor are carried out with the help of ESP tool algorithm. Finally, the index system of object oriented technology in Nanjian County was determined as follows: the optimal segmentation scale of ETM image was 40, the parameters of shape factor and compactness factor were 0. 4 and 0. 6. (2) combined with Nanjian County 2000, 2014 Nanjian County land use status data. The precision of land cover classification based on object-oriented method is verified. The overall accuracy of GF-1 classification results is 88.1293. The total accuracy of GF-1 image classification is 0.8469. The overall accuracy is 0.6925, which further proves that GF-1 land cover classification is higher. Based on this, the author explored and established an object-oriented classification method of GF-1 number with strong practicability in county area. (3) based on the analysis of land cover dynamic change information from 2000 to 2014, the results showed that the land cover types in Nanjian County in 2000 and 2014 were mainly woodland. They accounted for 46.36% and 57.71% of the total land use area respectively, followed by grassland, cultivated land and construction land, unused land and water land, and woodland, with the largest increase of 11.35% and the largest decrease of 11.95% in unused land. But it was not obvious; the comprehensive index of land cover degree increased from 211 to 225.91, increasing at the rate of 1.065% per year. All types of land cover had the phenomenon of turning in and out, the area of unused land was the largest, and the forest land was the least. The largest area is the water area, the least is cultivated land, the conversion area is small, relatively stable overall. On this basis, the problems of land cover change in Nanjian County from 2000 to 2014 were analyzed, such as woodland, increase of building land and water area, decrease of cultivated land and grassland area, low level of land use, unreasonable structure and layout of land use, etc. Some suggestions on land use planning, such as overall planning of urban and rural development, strengthening the utilization efficiency of water resources, developing economic forest industry, giving priority to forestry structure, increasing the protection of cultivated land and continuing to implement the policy of returning farmland to forest and grassland, are put forward.
【学位授予单位】：云南师范大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：P237;P208

【相似文献】