【作者】 宋莎；

【导师】 李廷轩； 张显峰；

【作者基本信息】 四川农业大学 ， 土壤学， 2010， 硕士

【摘要】 植被覆盖度是反映地表覆被状况的综合量化指标,在生态系统研究中起着十分重要的作用,植被覆盖度的研究在各个领域都有广泛应用。传统的植被覆盖度测量法只能用于评价小区域范围内生态环境的状况,随着遥感信息技术的快速发展,采用遥感技术反演地表信息已成为各行各业发展的需要,应用遥感方法反演植被覆盖度就是其中重要应用之一。本研究基于不同分辨率遥感数据和不同的方法建立了植被覆盖度的遥感反演模型,并对新疆石河子地区植被覆盖度状况进行了分析,探讨了影响植被覆盖变化的主要因素,得到研究结论如下：(1)数码相机照相法可以快速有效的获取植被覆盖度地面数据,一方面为反演模型的建立提供了基础数据,可提供建立模型的重要参数；另一方面,实测数据又可以作为验证数据用于评价反演模型的精度。在相片分类处理中,农作物和高、中覆盖度草地类型的分类精度最高,对整个反演结果误差的影响较小,而低覆盖度草地的分类精度最低,使得反演模型中该类型的反演结果误差最大。(2)不同分辨率遥感数据在建立反演模型时,由于地面实测数据“以点代面”的区域范围大小存在差异,因此导致各分辨率数据反演结果精度不同。对比TM与MODIS两种遥感传感器数据及两种反演方法的结果可以看出,基于二者的统计反演模型都能得到比较满意的反演结果,但总体上来说TM数据反演植被覆盖度的精度略高于MODIS数据经验反演模型精度,但MODIS数据的时间分辨率更高,因此运用基于MODIS数据的回归统计模型进行研究区植被覆盖度动态变化监测研究,既能满足反演的可行性,又能在大尺度上探讨植被覆盖度对生态环境系统的影响。(3)本研究在确定了适合用于植被覆盖度动态监测的最佳模型后,对2001年-2009年研究区的植被覆盖度结果进行了分析。结果表明,研究区植被覆盖度面积变化的总体趋势是从低覆盖度向高覆盖度逐渐转化,特别在2007年和2008年,高覆盖度面积出现了较大的增加,这说明研究区植被生态恢复情况较好,有利于整个生态环境的良性发展。影响植被覆盖度变化的主要自然因素包括土地利用／覆被类型的变化,区域年降水量的大小以及年平均气温的高低,其中高覆盖度面积的增加与耕地的扩张密切相关,而低覆盖度面积的变化与降水量有关。更多还原

【Abstract】 Fractional vegetation cover (FVC) is one of the comprehensive indicators to reflect the vegetation coverage on ground, and in the study of ecological system, it is very important. The previous methods of measuring vegetation coverage only can derive the information of a small area, and are hard to retrieve vegetation coverage in a large-scale area. Thus, the use of remote sense technology is necessary and effective in the study of vegetation dynamics. In the study, the retrieval model of fractional vegetation coverage was examined first, and then used to derive FVC information of the study area using the Landsat TM and MODIS data sets. The driven factors of the vegetation dynamics in the study area was also analyzed using some ancillary data.The vegetation coverage information was collected first using the digital camera and the photographs were processed in the ERDAS software application. The in-site FVC data were then used to derive the parameters of the inversion model, and some of the sample data points were also used to validate the model results. In the classification of the digital camera photos, more accurate FVC of vegetation plots were obtained in the crop land, high coverage grasslan, middle coverage grassland, and a lower results were got in the sparsely covered grassland due to the difficulty of plot design.While using different resolution remote sensing data to build inversion models, the ground measured data which will lead to the FVC results with different accuracy. Comparing the two kinds of remote sensing data and two inversion methods, the accuracy of inversion FVC method using TM data is higher than the one using MODIS data, and the lowest accuracy method is the dimidiate pixel model one using MODIS data. To monitor the vegetation cover dynamic change of the study area with the regression model using MODIS data, it can not only fit the feasibility of the inversion, but also discuss the effect between the FVC and the ecologic environment system in big scale.Using the inversion modle, we analyzed the changes of FVC from the year 2001 to 2009.It shows that the general tendency of the FVC area is from the low coverage to the high coverage, especially in 2007 and 2008 with a significantly increase of the high coverage area. This suggests that the vegetation of the study area is in good condition which is helpful to the sound progress of the whole ecological environment. The main natural factors of the vegetation cover changes are the changes of land use and cover type, the amount of the annual regional precipitation, and the level of the annual average temperature. The increase of the high coverage area is closely related to the expansion of cultivated land while the changes of the low coverage area are related to the precipitation.更多还原