针对用归一化差值植被指数（Normalized Difference Vegetation Index，NDVI）估算植被叶面积指数（Leaf Area Index，LAI）不仅需要大量地面 LAI 观测及其数据统计，且在植被 NDVI 饱和时难以估算 LAI 等问题，提出了一种基于数据挖掘技术的 LAI 遥感估算方法。该方法借助数据挖掘技术从有限的数据中挖掘和发现有用的信息，排除人为干扰，提高模型构建效率和精度。文中以安徽滁州地区杨树林为研究对象，获取研究区杨树林展叶期和花果期的 HJ-CDD 遥感影像，利用 LAI-2000同步测量杨树林 LAI；借助数据挖掘技术并基于杨树林展叶期和花果期估算的 LAI 值，通过筛选优化构建了杨树林生长过程中叶面积稳定期的 LAI 估算模型，并结合叶面积稳定期实测的 LAI 值验证表明该模型用于杨树林叶面积稳定期 LAI 估算的可靠性，为植被 NDVI 饱和时的LAI 遥感估算提供了一种有效的思路和方法。%As plenty of leaf area index (LAI)observation data are needed when vegetation normalized difference vegetation index (NDVI)is widely used to estimate LAI.In addition,it is difficult to estimate LAI when NDVI is saturated.A method that estimating LAI by data mining technology was proposed.This method can mine and discover useful information from limited data,eliminate human interference,and improve the efficiency and accuracy of modeling.In this paper,the aspen in Chuzhou city,Anhui province was taken as research object.HJ-CDD remote sensing images for aspen leaf production period and flowering and fruit-bearing period about the study area were obtained,and the aspen LAI was measured simultaneously with LAI-2000 .Applying the data mining technique,a novel model is established to estimate the LAI of aspen in the leaf constant period based on those in the leaf expansion period and flowering and fruit-bearing periods.Compared with the measured data of leaf constant period,this model is proved to be reliable and effective for vegetation LAI remote sensing estimation when NDVI was saturated.
展开▼
