*2.2. Cultivation of Samples*

Experimental samples of the tomato variety "Cooperative 906" were cultivated in the Venlo greenhouse of Jiangsu University. A 32 × 56 cm rectangular black plastic plate was used for raising the seedlings. Peat, perlite, and vermiculite were mixed to comprise the cultivation substrate, and the seeds were sown in the seedling tray. After budding, the seeds were transplanted into a flowerpot with a diameter of 23.8 cm and a height of 35 cm, and then cultivated by soilless potting using perlite as the substrate nutrient solution.

To eliminate interference, a standard concentration of Yamazaki nutrient solution was used to provide the same nutrients for the samples. High temperature and humidity conditions characterized the greenhouse to allow for the development of tomato leaf mildew. After 15 days of infection with tomato leaf mold, the mold was collected from lesioned areas. After collection, the mold was immediately placed into fresh-keeping sealed bags and placed in a portable refrigerated incubator in order to prevent evaporation and minimize the impact of external conditions. Finally, 240 effective samples were obtained, including those obtained from 40 healthy leaves and those obtained from 200 infected leaves. All tomato leaf samples were divided into four disease grades according to GB/T 17980.26-2000. Pictures of these four tomato leaf mildew grades are shown in Figure 2.

**Figure 2.** Tomato leaf mildew grades. (**a**) Level 0, (**b**) level 1, (**c**) level 3, (**d**) level 5.

Statistics of the effective sample sizes are shown in Table 1. We randomly arranged the samples of the four different tomato leaf mildew grades and randomly divided the training set and prediction set in a proportion of 2:1.

**Table 1.** Statistics of effective sample sizes.

