Premium
Field mapping of chemical composition of forage using hyperspectral imaging in a grass meadow
Author(s) -
Suzuki Yumiko,
Tanaka Katsuyuki,
Kato Wataru,
Okamoto Hiroshi,
Kataoka Takashi,
Shimada Hiroshi,
Sugiura Toshihiro,
Shima Eikichi
Publication year - 2008
Publication title -
grassland science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.388
H-Index - 19
eISSN - 1744-697X
pISSN - 1744-6961
DOI - 10.1111/j.1744-697x.2008.00122.x
Subject(s) - hyperspectral imaging , forage , calibration , environmental science , remote sensing , field trial , field (mathematics) , mathematics , statistics , agronomy , biology , geography , pure mathematics
Spatial information on the quantity and quality of forage is important for practicing precision agriculture in grasslands. The objective of this research was to generate a field map of the concentrations of total digestible nutrients (TDN) and crude protein (CP) of forage in an orchardgrass‐dominated meadow field (0.6 ha) using a hyperspectral imaging sensor. The sensor was mounted on the roof of a vehicle in order to scan the whole meadow field. Calibration models to estimate the concentrations of the chemical components were developed using a dataset consisting of observed concentrations and spectral data acquired from the hyperspectral images. The models were developed by multiple linear regression analysis with stepwise variable selection, and were evaluated by cross‐validation and evaluation index (EI) methods. The cross‐validation for the calibration model using 10 explanatory variables resulted in R 2 values of 0.74 and 0.76 for TDN and CP, respectively. Standard error of prediction (SEP) was, respectively, 4.51 and 1.04. Bias was close to 0 for both TDN and CP. The EI values ranged 28.1–29.4, showing rank C. These results show that the calibration models are practical for estimating the concentrations of TDN and CP. Finally, field maps of the concentrations of TDN and CP were generated by applying the calibration models to field‐scale image data. The maps were able to identify the variations in the concentrations of TDN and CP throughout the field. In conclusion, this system is a useful technique for obtaining maps of forage chemical composition for site‐specific field management.