Assessing nitrogen content of golf course turfgrass clippings using spectral reflectance
2004
Keskin, M. | Dodd, R.B. | Han, Y.J. | Khalilian, A.
Feasibility of a practical indoor reflectance-based sensor was studied to assess nitrogen content of turfgrass clippings from spectral reflectance data. Turfgrass clipping samples were obtained from commercial golf course putting greens, and their reflectance were measured using a dual-type spectroradiometer under artificial illumination 3 h and 51 h after mowing. The reflectance values in green band (520 to 580 nm) and the NIR band (770 to 1050 nm) increased as the nitrogen content increased. Four wavelength bands at 550, 680, 770, and 810 nm were selected to develop and compare several regression models with varying number of variables. All models performed well (R2 > 0.82) and predicted the nitrogen content with reasonable standard error of prediction (SEP) values (SEP < 0.62%) for the data taken 3 h after mowing. However, the data taken 51 h after mowing on the same samples did not yield good results (R2 < 0.60; SEP > 1.04%). A discriminant analysis showed that the regression model with two wavelength variables performed as well as the models with a higher number of variables. A simple reflectance sensor using even only one photodiode and bandpass filter can be developed to predict or classify the nitrogen content of turfgrass clippings when the reflectance data are taken within several hours of mowing.
اظهر المزيد [+] اقل [-]الكلمات المفتاحية الخاصة بالمكنز الزراعي (أجروفوك)
المعلومات البيبليوغرافية
تم تزويد هذا السجل من قبل National Agricultural Library