Non-Destructive Assessment of Dry Matter Yield, Nutrient Content and Feeding Value of Grass Swards with Imaging Spectroscopy

Abstract

The objective of this paper was to evaluate the accuracy of imaging spectroscopy for assessment of dry matter (DM) yield, nutrient content (N, P, K) and feeding value (sugar, ash and crude fibre content, neutral detergent fibre (NDF), acid detergent fibre and lignin (ADF and ADL) and in vitro digestibility) of heterogeneous swards typical for wet peat soils and ryegrass (Lolium perenne)-dominated grass swards on sand and clay soils in a field in the Netherlands. A mobile instrument was used that measures crop reflection from 1 m distance with a 2D camera and hyperspectral image-lines, covering a spectral range from 439-1680 nm with a spatial resolution of 0.2-0.6 mm2. Images were recorded just before harvest on various dates in experiments where sward quality, N or P application and urine application varied. Nutrient contents and feeding value of harvested material were chemically determined in the laboratory. Partial least squares models were calibrated and validated using classified images for peaty and ryegrass datasets. Root mean square errors of prediction (RMSEP) for DM yield were 390 and 479 kg/ha for ryegrass swards and swards on peat, respectively. RMSEP values were between 0.25-0.33%, 0.4-0.5 g/kg and 2.7-3.0 g/kg for N, P and K, respectively. Relative errors to the mean were between 9-14% for sugar content and between 4-6% for crude fibre, NDF and ADF content and in vitro digestibility. This study confirms earlier results on mini-swards. It is concluded that imaging spectroscopy provides robust and accurate means for assessment of DM yield, nutrient content and feeding quality of standing grass in the field, under various field conditions.