The aim of this work was to investigate the suitability of optical non-destructive methods such as hyperspectral imaging (HSI) and optical coherence tomography (OCT) for detection of P. expansum growth on apples. Treated and control apple samples of 'Golden Delicious' and 'Evelina' cultivars were monitored with HSI and OCT during incubation until the occurrence of the mould. In order to monitor changes in mould's proteome, apples were sampled at selected times for differential scanning calorimetry (DSC) and gel electrophoresis analysis. After basic image processing we performed the classification of mould in an images using support vector machine (SVM) algorithm and thresholding method after spectral unmixing. SVM method performed better for the 'Golden Delicious' cultivar and the thresholding method performed better for the 'Evelina' cultivar. The most successful method for early detection of mould was OCT, which detected mould growth on both cultivars on first day of incubation. Mould growth was slower on the 'Evelina' cultivar. Gel electrophoresis has proved unsuccessful in monitoring mould's proteome. Using DSC, we were able to detect differences between mouldy and healthy fruit, particularly on the basis of endothermic peak at 66 °C.
|
