Fig. 1. Tomato leaves hyperspectral image acquisition diagram

Fig. 2. Comparisons of reflectance of different ROIs after inoculation
(a): Tomato leaf on the 6th day after infection; (b): Different ROIs selection;(c): Comparisons of reflectivity of different ROIs; (d): Comparisons of reflectivity of leaf on the 1st and 9th day after infection

Fig. 3. RGB image of leaves

Fig. 4. Changes of spectral reflectance of samples observed for 7 consecutive data

Fig. 5. Original series (a) and interrelated series (b)

Fig. 6. Symbolic aggregate approximation (SAX) method

Fig. 7. Symbolic Fourier approximation (SFA) method

Fig. 8. Hyperspectral curve of diseased and healthy leaves for 7 consecutive observations
(a): Consecutive 7-day reflectivity of infected leaf 1; (b): Consecutive 7-day reflectivity of infected leaf 2; (c): Consecutive 7-day reflectivity of healthy leaf 1; (b): Consecutive 7-day reflectivity of healthy leaf 2

Fig. 9. Recognition results of SDSS-SAX-SFA-MRF model

Fig. 10. Recognition results of MDSS-SAX-SFA-MRF model
(a): SFA multidimensional spectrum; (b): SAX multidimensional spectrum; (c): SAX+SFA multidimensional spectrum

Fig. 11. Comparison results of MDSS-SAX-SFA-MRF model and SDSS-SAX-SFA-MRF model

Table 1. Parameters of symbolic methods and weighted random forest model