Soybean LAI Estimation with In-situ Collected Hyperspectral Data Based on BP-neural Networks

In order to evaluate the precision of hyperspectral reflectance model for monitoring soybean leaf area index (LAI), soybean canopy reflectance data collected with ASD spectroradiometers (350~1050 nm), which were cultivated in water - fertilizer coupled control conditions, and soybean LAI were collected simultaneously with LI-COR LAI-2000. First, correlation between reflectance, derivative reflectance against soybean LAI were conducted; secondly, five vegetation indices with reflectance at band 801 nm and 670 nm were applied to regresses against soybean LAI; finally, ANN-BP established for soybean LAI estimation with changeable nodes in hidden layers. It were found that soybean canopy reflectance show a negative correlation with soybean LAI, while it show a positive correlation with soybean LAI in near infrared region. Reflectance derivative have an intimate relation with soybean LAI in blue, green and red edge spectral region, and got maximum correlation coefficient in red edge region. All five vegetation indices have intimate correlation with soybean LAI, with regression determination coefficient R² ranged in 0.84 and 0.88. ANN-BP model can greatly improve soybean LAI estimation accuracy. Determination coefficient (R² = 0.92) obtained with 2 nodes in hidden layers, however, R² still can be improved with nodes in hidden layers increasing, and R² = 0.96 with 8 nodes in hidden layers. Still, it should be noticed that without indecent phonological soybean data participate model establishing, ANN-BP model can improve estimation accuracy with large room, and determination coefficient (R² = 0.98) can be obtained with 8 nodes in hidden layers. By above analysis, it indicated that, ANN-BP model can be applied to in-situ collected hyperspectral data for vegetation LAI estimation with quite accurate prediction, and in the future, ANN-BP model still should be applied to hyperspectral data for other vegetation biophysical a- nd biochemical parameters estimation.