Nondestructive prediction of acetolactate synthase of oilseed rape leaves using visible/near-infrared spectroscopy and BP neural networks

A new acetolactate synthase (ALS)-inhibiting herbicide of pyribambenz-propyl (PP) was applied to oilseed rape leaves with different positions. ALS could reflect the growing states of oilseed rape. Visible and near-infrared (Vis/NIR) spectroscopy was investigated for fast and non-destructive determination of ALS in rapeseed leaves. Partial least squares (PLS) analysis was the calibration method with different spectral preprocessing methods. The best PLS models were obtained by first-derivative spectra for ALS, Simultaneously, certain latent variables were used as the inputs of back propagation neural networks (BPNN) models with sigmoid function. The results demonstrated that BPNN method outperformed PLS method. The correlation coefficient, RMSEP and bias in validation set by BPNN were 0.994, 2.460 and -1.536 for ALS, respectively. The results indicated that Vis/NIR spectroscopy combined with BPNN could be successfully applied for the determination of ALS of rapeseed leaves. The results would be helpful for further on field analysis of using Vis/NIR spectroscopy to monitor the growing states and biological properties of oilseed rape.