利用反射率光谱估算植被干物质氮含量

Estimation of Vegetation Dry Matter Nitrogen Content with Reflectivity Spectrum

为了排除植被含水量对氮素总含量的可能性干扰,以及用叶绿素含量来估算氮含量时忽略叶绿体外氮素影响的问题,提出了利用植被干物质来估算植被氮含量的算法。以Lopex’93数据库为数据来源,拟合构建了植被干物质氮素含量(%)与光谱反射率之间的模型关系,并与鲜植被进行对比分析,又利用光谱反射率的4种变化形式(一阶导数、二阶导数、倒数的对数和倒数的导数)来验证分析。结果表明,由差值植被指数DVI(R760-R1520)拟合的复合模型Y=17.787×241.456x和指数模型Y=17.787×e5.487x对植被干物质氮含量具有较好的估算精度,其拟合值和真实值的相关系数R2=0.901。而用鲜植被拟合的模型精度明显低于干物质模型,经过反射率4种参数的验证,也说明了由DVI构建的干物质氮含量估算模型可以得到较好的结果。

Taking Lopex’93 database as the data source,fit and construct the relationship model between the nitrogen content (%)of vegetation dry matter and the spectral reflectance.It verifies the analysis by comparing with fresh vegetation and using the four change forms of the spectral reflectance (first derivative,second derivative,the logarithm of reciprocal and the derivative of reciprocal).Results show that the composite model Y =17.787 ×241.456x and exponential model Y =17.787 × e5 .487x constructing by difference vegetation index DVI (R760-R1520 )have better estimation precision to nitrogen content of vegetation dry matter and the correlation coefficient between fitted value and real value is R 2 =0.901.The accuracy of fresh vegetation fitting model is obviously lower than the dry matter model,and the estimated model which is constructed by DVI and dry matter can get good results through the validation of the four parameters of reflectivity.