Detection and quantification of Fusarium-specific disease symptoms in maize ears by spectral imaging

In today's maize breeding, resistance against ear fusariosis is an important challenge. Breeders need reliable methods for quantifying Fusarium-specific disease symptoms on maize ears. 20 maize varieties were inoculated with conidia of two predominant pathogenic Fusarium species at the stage of female flowering. After harvest, the ears were cut vertically into halves. Spectral images were taken of the inner side of the halves. A supervised classification procedure was applied to quantify the areas with Fusarium-specific disease symptoms. Results were compared to the visual estimation of an expert. The figures from the image analysis were slightly higher than the expert's estimation but they may be used to predict the expert estimation with r2 = 0.81.