基于高光谱的湿地植被分类研究

Classification of Wetland Vegetation Based on Hyperspectral

光谱特征的选择对于湿地植被的识别精度和效率有直接的影响。本文以萨克拉门托-圣华金三角洲为研究区,基于Hy Map航空高光谱遥感影像数据,分析湿地植被的一阶微分和二阶微分光谱特征。在上述分析的基础上基于均值置信区间的波段选择法对一阶微分、二阶微分进行波段选择,根据获取的有效特征波段构建特征集,利用C5决策树分类算法产生规则集,并对实验区的湿地植被进行了分类研究。结果表明:湿地植被的一阶微分、二阶微分能够突出不同湿地植被光谱曲线在不同波段的增速不同,利用均值置信区间的波段选择法能够对特征波段起到降维效果,根据降维后的特征波段采用C5决策树分类算法,可以实现湿地植被在物种水平上的识别,并达到较好的分类精度。

Certain spectral characteristics have a direct impact on accuracy and efficiency of identifying the wetland vegetation. Sacramento-San Joaquin Delta was taken as the test area, based on HyMap aerial hyperspectral remote sensing image data, the first-derivative and the second-derivative spectral features of the wetland vegetation were analyzed.Based on the abovementioned analysis, two order derivative were used for the selection of bands, the feature set was established and rule set was generated by using C5 algorithm. The results showed that: the first-derivative and the second-derivative spectral features of the wetland vegetation can prominent that in different bands the growth rate of different spectral curves of different wetland vegetation is different, the band selection method of using the mean confidence interval can reduce the dimension of the feature band, based on the dimension of the feature band by using C5 algorithm could be effective in distinguishing wetland vegetation and allowing for species-level detection.