【作者】 刘丽娟；

【导师】 范文义；

【作者基本信息】 东北林业大学 ， 森林经理学， 2008， 硕士

【摘要】 土地荒漠化是目前全球共同关注的重大环境问题。杜尔伯特蒙古族自治县是黑龙江省惟一完整的沙区县,荒漠化类型以风蚀为主,4、5、9、10月为风沙活动最为频繁强烈的月份。受风沙危害,草场逐年退化。为遏制沙化土地和其他类型荒漠化土地扩展的趋势,迫切需要对其荒漠化现状作出正确评价,找出相应的治理措施,进而改善荒漠化地区的生态环境。本论文以Landsat 7 ETM+遥感影像为数据源,通过评价因子建立遥感信息模型对典型区杜尔伯特蒙古族自治县土地进行荒漠化现状评价。从植被状况、土壤状况、地表状况等角度确定荒漠化评价因子:植被覆盖度、生物量、裸沙占地百分比、土壤质地和地表结皮状况,应用3S技术及地面调查相结合的手段,建立研究区的荒漠化评价指标体系。主要研究结论:用线性光谱混合模型反演裸沙占地百分比,非密度模型反演植被覆盖度,并将缨帽变换的第1、2分量分别与反演结果进行相关性分析,表明模型反演精度较高;多元回归分析建立生物量预估模型反演生物量,决定系数R~2=0.70003,回归结果显著,并用10个实测数据进行模型的独立性检验,预测精度较高为79.0067%,满足本研究的需要;改进的基于IHS的小波变换融合方法,融合多光谱和全色影像,提取土壤质地和地表结皮状况训练区。通过主客观评价,融合后影像目视效果大大提高,增加了训练区提取的准确性;将五个评价因子反演结果建立荒漠化评价可视化遥感信息模型。此模型对研究区进行荒漠化程度定量评价,精度达88.75%;然后以栅格形式进入GIS并形成专题图,进行荒漠化面积统计。结果表明,轻度荒漠化占荒漠化面积的57.6%(其中水域占13.23%),中度荒漠化占21.77%、重度和极重度荒漠化占7.4%,这说明杜尔伯特蒙古族自治县荒漠化总体程度较轻,治理措施已见成效。更多还原

【Abstract】 Desertification is a serious environmental question at present. Duerbote mongolian nationality autonomous county is the only integrated desert county, the dominating type of desertification is wind-corroded. On April, May, September and October, sand and wind is frequent. Damaged by them, Meadow degraded year after year. For Keeping within limits the expand trend of desertification, correctly evaluats for deserted actuality, and makes the relevant measures, then improves the environment of desertification area are in urgent.The main researching conclusion and creative point are made from this paper:Liner spectral mixing model (LSMM) retrivals the occupation ratio of bare sand land, non-density model retrivals vegetation coverage, and then make correlation analysis with tasseled cap of the first and second variable. The result indicates the precision of retrival is good. Multiple regression analysis makes the biomass estimate model and the result is prominent, and R~2 is 0.70003. Then independent test is made. Predicted precision is 79.0067%, and meet this study need. Using improved wavelet transform fusion method based on IHS transform fuses multiple spectral and panchromatic images, and picks up the training area of soil texture and earth’s surface status. Through the subjective and objective evaluation, the eye interpretation effect of fused image improves greatly, and increases the accuracy of training choice. The retrival result of five factors establishes the visual RS information model. Using this model to evaluate, the precision is 88.75%.Then the evaluated results enter GIS in raster and make thematic map to calculate acreage. The result indicates: slight regions account for 57.6%(the water is 13.23%),middle regions are 21.77%, and severe and extreme regions occupy 7.4%.All these show the desertification extent of duerbote mongolian nationality autonomous county is slight in total, the father measures win initial success.更多还原