机载AISA EagleⅡ高光谱数据在温带天然林树种分类中的应用

以内蒙古根河地区的温带天然林试验区为研究对象，采用支持向量机(SVM)方法，对经过大气校正后的机载AISAEagleII高光谱地表反射率影像分航带进行树种分类，将地面光谱测量与高光谱同平台的高分辨率航空相片结合，进行训练样本的选择，使用地面样地数据对分类结果进行验证。结果表明:利用AISAEagleII高光谱影像对温带天然林区分类的总体精度和kappa系数分别达到了96．71%和0．95;灌木分类精度最高，其制图精度和用户精度分别达到了98．07%和98．31%;落叶松和白桦的用户精度分别为98%和94%。

Taking the temperate natural forest experimental area in Genhe, Inner Mongolia, China as the research object, this study employed the Support Vector Machine (SVM) approach to perform tree species classification on the atmospherically corrected airborne AISAEagle II hyperspectral surface reflectance imagery by flight strips. Training samples were selected by integrating field spectral measurements with high-resolution aerial photographs collected on the same hyperspectral platform, and the classification results were validated using field plot data. The results demonstrated that the overall accuracy and Kappa coefficient of tree species classification in the temperate natural forest using the AISAEagle II hyperspectral imagery were 96.71% and 0.95, respectively. Shrub exhibited the highest classification accuracy, with its mapping accuracy and user's accuracy reaching 98.07% and 98.31%, respectively. The user's accuracies of larch and white birch were 98% and 94%, respectively.

搜集汇总

背景与挑战

背景概述

该数据集包含机载AISA EagleⅡ高光谱数据在温带天然林树种分类中的应用研究，通过支持向量机方法实现了高精度的树种分类，总体精度和kappa系数分别达到96.71%和0.95。数据集属于国家重点研发计划项目，数据格式为文档，数据量为2.16 MB。