Phytopathogens pose a serious threat to agriculture, causing a decrease in yield and product quality. This necessitates the development of methods for early detection of phytopathogens, which will reduce losses and improve product quality by using lower quantities of agrochemicals. In this study, the efficiency of spectral imaging in the early detection and differentiation of diseases caused by pathogens of different types (Potato virus X (PVX) and the bacterium Pseudomonas syringae) was analyzed. An evaluation of the visual symptoms of diseases demonstrated the presence of pronounced symptoms in the case of bacterial infection and an almost complete absence of visual symptoms in the case of viral infection. P. syringae caused severe inhibition of photosynthetic activity in the infected leaf, while PVX did not have a pronounced effect on photosynthetic activity. Reflectance spectra of infected and healthy plants were detected in the range from 400 to 1000 nm using a hyperspectral camera, and the dynamics of infection-induced changes during disease progression were analyzed. P. syringae caused a strong increase in reflectance in the blue and red spectral ranges, as well as a decrease in the near-infrared range. PVX-induced changes in the reflectance spectrum had smaller amplitudes compared to P. syringae, and were localized mainly in the red edge (RE) range. The entire set of normalized reflectance indices (NRI) for the analyzed spectral range was calculated. The most sensitive NRIs to bacterial (NRI510/545, NRI510/850) and viral (NRI600/850, NRI700/850) infections were identified. The use of these indices makes it possible to detect the disease at an early stage. The study of the identified NRIs demonstrated the possibility of using the multispectral imaging method in early pathogen detection, which has high performance and a low cost of analysis.