Abstract

Leaf Area Index (LAI) is an important biophysical parameter and is a critical variable in many ecology models, productivity models, and carbon circulation studies. To assess and compare various hyperspectral models in terms of their prediction power of tobacco LAI, tobacco canopy hyperspectral reflectance data of the root extending stage, fast growing stage, and mature stage in different water-nitrogen conditions were collected with a FieldSpec HandHeld spectroradiometer. Based on the pot experiment data, an evaluation of tobacco LAI retrieval methods was conducted using four vegetation indices, principal component analysis (PCA), and neural network (NN) methods. The estimated effects of the three methods were then compared. Results indicated that all three methods have ideal effects on LAI estimation. Determination coefficients (R²) of the validated models of vegetation indices, PCA, and NN were (0.768 ~ 0.852), 0.938, 0.889, respectively. The PCA and NN methods show higher precision. The stability of the PCA validated model is the best because its Root Mean Square Error (RMSE) of 0.172 is smaller than those of  the vegetation indices (0.237 ~ 0.322) and NN (0.195). As a whole, the PCA and NN methods could improve the retrieval precision and were prior selection for LAI estimation.

Key words: Hyperspectral, Flue-cured Tobacco, LAI, vegetation Indices, principal component analysis, neural network.