| 引用本文: | 史晓艳,李维弟,余露,等.玛纳斯河流域农灌区土壤盐渍化遥感定量评价[J].灌溉排水学报,2018,37(11):69-75+83. |
| SHI Xiaoyan,LI Weidi,YU Lu,et al.玛纳斯河流域农灌区土壤盐渍化遥感定量评价[J].灌溉排水学报,2018,37(11):69-75+83. |
|
| 摘要: |
| 【目的】提高利用遥感影像解译盐渍化土壤的精度。【方法】以新疆玛纳斯河流域为研究区，利用Landsat 8 OLI遥感数据，采用主成分法分析了归一化盐分指数指标（NDSI）、地表反照率指标（Albedo）、亮度指标（Bright）、绿度指标（Green）、归一化植被指数指标（NDVI）、湿度指标（Wet）、蓝色波段指标（B）、红色波段指标（R）、中红外波段指标（SWIR1）和氧化铁指标（[IFe2O3]）10个指标之间的关系，构建了不同盐渍化程度土壤的函数表达式，结合实测盐分数值与不同程度盐渍化土壤栅格图的拟合，确定了盐渍化程度的判别阈值，叠加制作了研究区盐渍化土壤类型划分图。【结果】基于遥感数据和土壤实测样点划分出的土壤盐渍化程度分布图，轻度盐渍化土壤判别准确率为68.97%，中度盐渍化为76.47%，重度盐渍化为83.83%，平均判别准确率达76.42%，轻度盐渍化土壤判别准确率相对较低，重度和中度盐渍化土壤的分布区域有较好的一致性。【结论】将遥感数据与实测土壤盐分数据相结合，能够提高盐渍化土壤的判别精度。 |
| 关键词: 土壤； 盐渍化； 遥感； 主成分分析； 评价 |
| DOI：10.13522/j.cnki.ggps.20180134 |
| 分类号: |
| 基金项目: |
|
| Using Remote Sensing to Evaluate Soil Salinization Distribution Over the Irrigation Areas in the Manas River Basin |
|
SHI Xiaoyan, LI Weidi, YU Lu, WANG Haijiang*, SONG Jianghui, ZHU Yongqi
|
|
Department of Resources and Environmental Sciences, College of Agronomy, Shihezi University, Shihezi 832000, China
|
| Abstract: |
| 【Objective】Remote sensing has been increasingly used in agronomic management and in this paper, we studied the feasibility of using it to estimate saline soil distribution in an irrigation area.【Method】 We took Manas River basin in Xinjiang as an example. Landsat 8 OLI remote sensing data was used to analyze the relationship between NDSI (Normalized Difference Soil Index), Surface Albedo Index (Albedo), Bright, Green, Normalized Difference Vegetation Index (NDVI), Wet, Blue Band (B), Red Band (R), Middle Infrared Band (SWIR1) and Indexes of Ferric Oxide ([IFe2O3]), based on the principal analysis method. The results were then used to construct functional relationship between these indices for soils with different salinization degree. The threshold of soil salinity was determined comparing real salinity measurements and the raster map for soils with different salinization degree, from which the maps of salinized soil were created.【Result】 Comparing the measured data from soil samples and those calculated from the RS data revealed that the accuracy of the estimated soil salinity was approximately 76.42%, of which the accuracy for slightly salinized soil was 68.97%, moderate salinized soil was 76.47%, and severe salinized soil was 83.83%. 【Conclusion】 Combining RS data with measured soil salt content can improve the estimation of salinized soil in large river basins. |
| Key words: soil; salinization; remote sense; principle analysis; evaluate |
