基于NDVI时序数据的不同水源灌溉面积分类研究--以人民胜利渠灌区为例

灌溉面积是水土资源利用的重要指标,灌溉面积监测是灌区和农业水资源管理的基础。【目的】探索不同水源的灌溉面积的遥感分类方法。【方法】基于人民胜利渠灌区2017年的Landsat8卫星数据和2012年的MODIS卫星数据,提取了灌区NDVI时序数据曲线,分别采用监督分类和非监督分类方法对灌区不同水源的灌溉面积进行分类,并结合实地调查资料对分类结果进行验证。【结果】非监督分类方法并不适用于灌区尺度的灌溉水源分类;基于MODIS数据的分类效果较差,分类精度为57.14%;基于Landsat8数据的NDVI时序数据和监督分类的结果较好,分类精度达到73.58%。【结论】空间分辨率较高的NDVI时序数据可对灌区灌溉水源进行分类。

Using NDVI Time Series Data to Discern Areas Irrigated with Different Water Sources:Taking People’s Victory Canal Irrigation District as an Example

【Objective】Knowledge of spatiotemporal change in areas irrigated with different water sources in a region is essential for improving its soil and water management, and this paper presents and verifies a satellite imagery-based method to discriminate areas irrigated with surface and subsurface water with the People’s Victory Canal Irrigation District(PVCID) taken as an example. 【Method】 NDVI time series data of PVCID were extracted from the Landsat8 satellite imagery in 2017 and the MODIS satellite imagery in 2012, the areas in which irrigated with different water sources were identified using both supervised and unsupervised classification method respectively. The results obtained from the two methods were verified against ground-truth data measured from the field in the district. 【Result】The unsupervised classification method worked poorly. Comparison with field measurements showed that the accuracy of the supervised classification was 57.14% and 73.58% in using the MODIS data and the Landsat8 imagery data, respectively. 【Conclusion】NDVI time series imagery with high spatial resolution can be used to differentiate areas irrigated using different water sources at catchment and irrigation district scale.