基于遥感技术的草地信息提取及动态监测是草地资源调查和科学管理的重要手段。本研究以TM和HJ星CCD数据为研究数据，采用监督分类和非监督分类的混合分类方法，提取研究区草地资源信息，通过逻辑运算获取NDVI指数，并进行等级划分，以转移矩阵和栅格运算为主要方法分析草地资源的动态变化特征。研究结果表明：混合分类方法与监督分类相比，2000年和2010年分类总精度分别提高了8．90％和8．47％；2000-2010年草地面积减少了325．03km^2，宜农耕地面积增加了183．75km^2，居民点用地面积增加了3．27km^2。荒漠化面积进一步增加，达到852．59km。，新增149．45km^2；草地覆盖度整体呈下降趋势，二、三等级中高覆盖度草地面积分别减少了208．02km^2和242．97km^2，低覆盖度的五等级草地面积增加了406．87km^2。%Grassland information extraction and dynamic monitoring based on remote sensing is an important tool of grassland resources survey and scientific management. The study used the mixed method of supervised and unsupervised classification to extract grassland resources information of the study area based on TM and HJ Star CCD satellite data. After obtaining the NDVI index by Boolean calculation and making the classification, dynamic changing characteristics of grassland resources were analyzed with the method of transfer matrix and the grid operations. The results show that： compared with supervised classification in the year of 2000 and 2010, the total classification accuracy of hybrid classification method was increased by 8.90% and 8.47% respectively; from 2000-2010, grassland area was decreased by 32503ha, agricultural land area was increased by 18375ha, residential land area was increased by 327ha and desert area was further increased to 85259ha, increased 14945ha; the overall of grass coverage degree has a downward trend. The area of high respectively. Key words： -coverage grassland of the second and third grade was decreased by 20802ha and 24297ha The area of low coverage of the fifth grade had an increase of 40687ha.
展开▼
