基于无人机高光谱的烟田涝灾早期识别

目的]为准确、及时估算烟田涝灾损失程度.方法]以搭载高光谱成像仪的无人机采集强降水后的烟田高光谱遥感影像,对影像进行图像分割、几何校正、辐射校正、地表反射率反演等处理,根据烟株倒伏程度将烟田分为受灾烟田、正常烟田和土壤3个类别,并构建兴趣区(region of interests,ROI),采用光谱角匹配算法对涝灾...

Early identification of tobacco field waterlogging disaster based on UAV hyperspectral Images

To realize accurate and quick estimation of tobacco field waterlogging disaster, Unmanned Aerial Vehicle (UVA) equipped with hyperspectral imager was used to collect remote sensing images of the study area, and then segmentation, geometric correction, radiometric correction, surface reflectance inversion of collected images were carried out.. Based on the degree of tobacco plant lodging tobacco field were classified into three categories: the affected fields, normal fields, and bared soil, and ROIs (Region of interests) were built based on the results. Spectral angle matching algorithm was used to extract and classify the waterlogging area from hyperspectral image of tobacco field. The spectral correlation coefficient and spectral angle were used to evaluate the matching degree between ROI and spectral curve. The Confusion matrix was used to evaluate the accuracy of the classification results, withoverall classification accuracy reaching 91.8% and Kappa coefficient reaching 0.85. The results shows that the UAV hyperspectral information comvbined with spectral angle matching algorithm can effectively identify waterlogging area in tobacco fields, which provides technical support for quick estimation of tobacco filed waterlogging.