Measurement of soil total nitrogen (STN) is urgently important concerning requirement of precise and quantitative fertilizer application. To overcome the shortage of routine chemical detection, vis- short wavelength near infrared spectroscopy (Vis-SWNIRS) was employed as an accurate, cheap and timely alternative. Kennard-Stone algorithm was utilized for sample set partitioning, where 22 of the 32 samples were selected as calibration set, and the remaining 10 were included in validation set. No outliers were discovered under criterion based on spectral Mahalanobis distance and Dixon testing. Partial least squares regression (PLSR) was used to build STN detection model based on Vis-SWNIRS with cross validation by leave-one-out method. As a result, 5 principal components turned to be optimal considering model performance, where correlation coefficients for calibration and validation were 0.9724 and 0.8691, respectively, with PRESS (Prediction Residual Error Sum of Square) of 0.0684. It is feasible to detect STN by Vis-SWNIR spectroscopy.
Weitere Kapitel dieses Buchs durch Wischen aufrufen
Bitte loggen Sie sich ein, um Zugang zu diesem Inhalt zu erhalten
Sie möchten Zugang zu diesem Inhalt erhalten? Dann informieren Sie sich jetzt über unsere Produkte:
- Detection of Soil Total Nitrogen by Vis-SWNIR Spectroscopy
- Springer Berlin Heidelberg
Neuer Inhalt/© ITandMEDIA