High early biomass production has been suggested as a useful selection criterion for improving cereal yields in mediterranean environments. Current methods for measuring biomass production in cereal plots involves destructive sampling which is not suitable for routine use by plant breeders where large numbers of samples are to be screened. The measurement of spectral reflectance using ground-based remote sensing techniques has the potential to provide a nondestructive estimate of plant biomass production. This field study investigated the relationship between spectral reflectance and early biomass production in barley and wheat and assessed the feasibility of using spectral reflectance techniques in cereal-breeding programs. Spectral reflectance measurements (0.40-2.5 �m wavelengths) were made before stem elongation for barley and wheat grown in separate experiments at Wongan Hills, Western Australia. Cereal biomass was closely associated with spectral reflectance in the visible (0.44-0.66 �m) and mid infrared (1.50-1.80 �m and 2.00-2.40 �m) regions, but not in the near infrared (0.71-1.10 �m). The poor correlation between reflectance in the near infrared region and biomass production was due to the low amounts of biomass present and the high reflectivity of the sandplain soils at these wavelengths. The close associations in the visible and mid infrared regions provides promise for use in estimation of cereal biomass, but further research is required to refine the technique for use in large scale breeding programs.
Visible to near-infrared multispectral images dataset for image sensors design
