论文部分内容阅读
土壤盐渍化是干旱、半干旱农业区主要的土地退化问题,同时也是一个重要的环境问题。选取塔里木盆地南缘克里雅河流域绿洲为研究靶区,利用Landsat-ETM+卫星图像数据和野外调查数据分析盐渍化土壤与地表反照率(Albedo)、土壤盐分指数(SI)之间的关系。回归分析发现,盐渍化土壤在SI-Albedo特征空间分布具有显著规律,即非盐渍化土壤呈团状分布;轻、中度盐渍化土壤具有线性分布特征;非盐渍化土壤与轻度盐渍化土壤分异明显。结合分异规律,编制分类算法模型,得到研究区盐渍化土壤信息提取结果,并与传统监督最大似然分类法结果进行对比分析。结果表明,在SI-Albedo特征空间中定量快速提取盐渍化土壤信息的总体效果较好,对准确且自动提取干旱区盐渍化土壤信息以及区域尺度盐渍化遥感定量监测研究具有重要意义。
Soil salinization is a major land degradation problem in arid and semi-arid agricultural areas, and it is also an important environmental issue. Using the Oasis in the Keriya River basin in the southern Tarim Basin as the target, the relationship between salted soil and albedo and soil salinity index (SI) was analyzed using Landsat-ETM + satellite imagery and field survey data . Regression analysis showed that salinized soils had a significant spatial distribution in the SI-Albedo feature, that is, the non-salinized soils distributed in clusters. The light and moderate salinized soils had a linear distribution, while the non-salinized soils and the light Degree salinization soil obvious difference. Combined with the law of differentiation, the classification algorithm model was compiled, and the results of salinization soil information extraction were obtained and compared with the results of the traditional supervised maximum likelihood classification. The results show that the overall effect of quantitative and rapid extraction of salinized soil information in SI-Albedo feature space is good, and it is of great significance to accurately and automatically extract salinized soil information in arid areas and quantitatively monitor the remote sensing of regional scale salinization.