Predicting the quality of ryegrass using hyperspectral imaging

Area of Science:

• Agricultural Science
• Plant Science
• Remote Sensing

Background:

• Forage quality is vital for animal performance and a key limitation in pasture-based systems.
• Improving sugar, lipid, protein, and energy content in forage plants is essential.
• Current methods for assessing forage composition are often expensive and destructive.

Purpose of the Study:

• To evaluate the potential of hyperspectral imaging (HSI) for non-invasive assessment of forage chemical composition.
• To determine the accuracy of HSI in predicting various quality attributes in perennial ryegrass (Lolium perenne).

Main Methods:

• Hyperspectral image data were collected from 185 perennial ryegrass accessions across 550-1700 nm wavelengths.
• Collected spectral data were correlated with 13 analyzed forage quality attributes.
• Statistical models were developed to predict forage composition from HSI data.

Main Results:

• HSI models demonstrated medium to high predictive power for total sugars (R²=0.58), high molecular weight sugars (R²=0.63), %Ash (R²=0.50), and %Nitrogen (R²=0.70).
• Significant models were also developed for other attributes including low molecular weight sugars, NDF, ADF, DOMD, ME, DM, Ca, and OM.
• HSI effectively differentiated chemical composition between pseudostems and leaves within plants.

Conclusions:

• Hyperspectral imaging technology shows promise for in-field estimation of forage composition in perennial ryegrass.
• HSI provides a cheaper, non-destructive, high-throughput screening tool for genetic selection and breeding.
• This technology can accelerate the development of improved forage varieties with enhanced nutritional value.