基于多时相高分一号影像的玉米涝灾监测

结合多时相遥感数据获取的对玉米涝灾的空间分布和程度进行快速大范围提取,可为确定农作物的受灾范围、作物受损评估提供参考依据。通过获取受涝玉米区灾前、灾中、灾后的高分一号影像,结合多时相的归一化水体指数(NDWI)变化分析,设置合适的阈值,用决策树的方法获取玉米受涝范围。根据拔节期玉米受涝前后的归一化植被指数(NDVI)变化特征,提出一种利用综合多时相NDVI差值的指数ST|ΔD|来监测玉米受涝程度。根据实际野外采样点验证及产量测点分析,受涝范围的提取精度达到87%,受涝程度划分精度最高达到82%。

Monitoring of Maize Flooding based onMulti - temporal GF-1

Based on multi-temporal remote sensing data, the spatial distribution and degree of waterlogging in maize are extracted rapidly and widely, which can provide a reference for determining the damage range of crops and assessing crop damage.By acquiring the gaofen-1 images before, during and after the waterlogging maize disaster, and combining with the multi-temporal NDWI change analysis, the appropriate threshold value was set, and the waterlogging range of maize was obtained by the decision tree method. According to the variation characteristics of NDVI before and after waterlogging of maize at jointing stage, an index ST|?D| integrating multi-temporal NDVI difference value is put forward to monitor the waterlogging degree of maize. According to the verification of actual field sampling points and analysis of production test points, the extraction accuracy of waterlogging range reaches 87%, and the classification accuracy of waterlogging degree reaches 82%.