Advances carried out in the last 20 years in imaging spectroscopy and thermal imaging for the retrieval of plant traits in the context of biotic and abiotic stress detection will be discussed. In particular, new developments in sensor miniaturization and physical modelling techniques have enabled a fast progress in the early detection of biotic and abiotic stress as part of precision agriculture efforts. The radiative-transfer based methodologies required for the pre-visual detection of stress using manned and unmanned vehicles carrying hyperspectral and thermal cameras will be described for the particular case of Xylella fastidiosa (Xf) outbreak in Europe. Results obtained using high?resolution hyperspectral and thermal airborne cameras revealed subtle spectral changes related to the Xf infection before the symptoms were visible in the field. We found that spectral plant traits calculated from the blue region, solar-induced chlorophyll fluorescence and remotely sensed tree crown temperature were the most statistically-significant indicators for the pre?visual detection of Xf symptoms. The use of these physiology-based spectral plant traits along with machine- and deep-learning algorithms will be described, discussing the existing gaps for the operational use of these surveillance methods at larger scales.