以土地利用为基础的多伦县沙质荒漠化评价遥感信息模型

Remote sensing information model for sandy desertification assessment in Duolun based on land use

从土地利用的角度对荒漠化进行分类与评价,建立以土地利用类型为基础的荒漠化评价指标体系。利用高分辨率卫星影像修正线性光谱混合分解模型分解的TM影像植被分量,建立适于提取干旱半干旱地区植被覆盖度的模型来提取植被信息;利用线性光谱混合分析模型计算裸沙占地百分比;用实测数据与热惯量进行拟合,按不同的土地利用类型分别建立拟合方程反演土壤含水率;引入遥感代用指标获得土壤氧化铁含量;采用目视解译获得土壤质地的编码图并进行影像化。获取每个像元全部评价因子的指标值,在现有荒漠化评价方法的基础上,建立以像元为单位的荒漠化评价定量化遥感信息模型并与高尚武模型沙漠化评价结果进行对比。结果表明:建立的评价模型总体精度达90.3%,较高尚武模型提高21.6%,尤其在评价耕地时,精度提高35%,体现了分土地利用类型进行评价的科学性。所建立的遥感信息模型评价土地荒漠化是可行的。

The desertification was classified and evaluated based on the land use point of view and the evaluation index system based on land use was set up.TM image was decomposed by linear spectral mixture mode and the vegetation component of TM image was amended by the high-resolution satellite image.Then the model,which was fit for the extraction of the arid zones＇ vegetation coverage,was set up to extract the information of vegetation.The occupation ratio of bare sandy land was obtained using linear mixed spectrum model.Soil moisture was retrieved by thermal inertia according to land use types.The substitute index of remote sensing was used to obtain the content of ferric oxide.And the code of soil texture was obtained by view interpretation.All the values of evaluative factors in every pixel could be acquired.Based on the current desertification assessment method,quantitative remote sensing information model on the basis of the pixel was developed.The results showed that the overall precision of this mold was 90.3%,and it was improved by 21.6% compared with the mold of Gao Shang-wu.The precision was improved by 35% on the evaluation of farmland.All of those showed that it was scientific to evaluate desertification from the land use point of view.By this token,it is feasible to evaluate desertification using quantitative remote sensing information model of this paper.