摘要
以山东省拔节初期至抽穗末期小麦为研究对象,利用高光谱分析技术及Excel、DPS统计分析软件分析了32组光谱数据与小麦株高的相关性,以期探索利用MOD IS数据估算小麦株高的方法。结果表明,光谱反射率与株高在红光b1、短近红外b2、蓝光b3、绿光b4及长近红外b7共5个波段的相关性达到了极显著水平,其中,可见光b1、b3、b4及长近红外b7与株高负相关,短近红外b2与株高正相关,用红光b1、短近红外b2建立的b2/b1、b2-b1、(b2-b1)/(b2+b1)3个植被指数,与小麦株高均呈极显著正相关关系,长近红外b7反射率与小麦株高相关度最高,相关系数为-0.871。分别用反射率、植被指数建立8个估算模型,相对误差最高仅为13.7%,模型的准确度较高,因此可以利用MOD IS数据进行小麦株高估算。
In order to study the method for estimating wheat plant height using MODIS data set, with the wheat from early -jointing stage to last -heading stage in Shandong Province as materials, the correlation between spectrum data and wheat plant height were analyzed by hyper spectrum analysis method and Excel and DPS statistical software. The results showed that the correlation between spectral reflectance and plant height reached significantly level in red band bl, short infrared band b2, blue band b3, green band b4 and long infrared band b7, which was negative in visible light bands and long infrared band but positive in short infrared band. Three vegetation indices were set with red band bl and short infrared band b2, which were b2/b1, b2 -b1 and （b2- bl ）/（ b2 ＋ b1）. The three indices showed significantly positive correlation with wheat plant height. Among these spectrum characters, long infrared band b7 had the highest correlation with plant height, whose correlation coefficient was -0. 871. Eight estimating models with higher accuracy were set with reflectance and vegetation indices, and the highest RMSE was 13.7%. So wheat plant could be estimated by MO- DIS data set.
出处
《山东农业科学》
2009年第2期5-7,11,共4页
Shandong Agricultural Sciences
基金
山东省农业科学院2006YCX033
2006YQN047
2007YCX026项目支持
