Fungal infections trigger defense or signaling responses in plants, leading to various changes in plant metabolites. The changes in metabolites, for example chlorophyll or flavonoids, have long been detectable using time-consuming destructive analytical methods including high-performance liquid chromatography or photometric determination. Recent plant phenotyping studies have revealed that hyperspectral imaging (HSI) in the UV-range can be used to link spectral changes with changes in plant metabolites. To compare established destructive analytical methods with new non-destructive hyperspectral measurements, the interaction between sugar beet leaves and the pathogens Cercospora beticola, which causes Cercospora leaf spot disease (CLS), and Uromyces betae, which causes sugar beet rust (BR), was investigated. With the help of destructive analyses, we showed that both diseases have different effects on chlorophylls, carotenoids, flavonoids, and several phenols. Non-destructive hyperspectral measurements in the UV-range revealed different effects of CLS and BR on plant metabolites resulting in distinct reflectance patterns. Both diseases resulted in specific spectral changes that allowed differentiation between the two diseases. Machine learning algorithms enabled the differentiation between the symptom classes and recognition of the two sugar beet diseases. Feature importance analysis identified specific wavelengths important to the classification, highlighting the utility of the UV-range. The study demonstrates that HSI in the UV-range is a promising, non-destructive tool to investigate the influence of plant diseases on plant physiology and biochemistry.

Over the last 20 years, researchers in the field of digital plant pathology have chased the goal to implement sensors, machine learning and new technologies into knowledge-based methods for plant phenotyping and plant protection. However, the application of swiftly developing technologies has posed many challenges. Greenhouse and field applications are complex and differ in their study design requirements. Selecting a sensor type (e.g., thermography or hyperspectral imaging), sensor platform (e.g., rovers, unmanned aerial vehicles, or satellites), and the problem-specific spatial and temporal scale adds to the challenge as all pathosystems are unique and differ in their interactions and symptoms, or lack thereof. Adding host-pathogen-environment interactions across time and space increases the complexity even further. Large data sets are necessary to enable a deeper understanding of these interactions. Therefore, modern machine learning methods are developed to realize the fast data analysis of such complex data sets. This reduces not only human effort but also enables an objective data perusal. Especially deep learning approaches show a high potential to identify probable cohesive parameters during plant-pathogen-environment interactions. Unfortunately, the performance and reliability of developed methods are often doubted by the potential user. Gaining their trust is thus needed for real field applications. Linking biological causes to machine learning features and a clear communication, even for non-experts of such results, is a crucial task that will bridge the gap between theory and praxis of a newly developed application. Therefore, we suggest a global connection of experts and data as the basis for defining a common and goal-oriented research roadmap. Such high interconnectivity will likely increase the chances of swift, successful progress in research and practice. A coordination within international excellence clusters will be useful to reduce redundancy of research while supporting the creation and progress of complementary research. With this review, we would like to discuss past research, achievements, as well as recurring and new challenges. Having such a retrospect available, we will attempt to reveal future challenges and provide a possible direction elevating the next decade of research in digital plant pathology.

In recent studies, the potential of hyperspectral sensors for the analysis of plant-pathogen interactions was expanded to the ultraviolet range (UV; 200-380 nm) to monitor stress processes in plants. A hyperspectral imaging set-up was established to highlight the influence of early plant-pathogen interactions on secondary plant metabolites. In this study, the plant-pathogen interactions of three different barley lines inoculated with Blumeria graminis f. sp. hordei (Bgh, powdery mildew) were investigated. One susceptible genotype (cv. Ingrid, wild type) and two resistant genotypes (Pallas 01, Mla1- and Mla12-based resistance and Pallas 22, mlo5-based resistance) were used. During the first 5 days after inoculation (dai) the plant reflectance patterns were recorded and plant metabolites relevant in host-pathogen interactions were studied in parallel. Hyperspectral measurements in the UV range revealed that a differentiation between barley genotypes inoculated with Bgh is possible, and distinct reflectance patterns were recorded for each genotype. The extracted and analysed pigments and flavonoids correlated with the spectral data recorded. A classification of noninoculated and inoculated samples with deep learning revealed that a high performance can be achieved with self-attention networks. The subsequent feature importance identified wavelengths as the most important for the classification, and these were linked to pigments and flavonoids. Hyperspectral imaging in the UV range allows the characterization of different resistance reactions, can be linked to changes in secondary plant metabolites, and has the advantage of being a non-invasive method. It therefore enables a greater understanding of plant reactions to biotic stress, as well as resistance reactions.

Deep neural networks have shown excellent performances in many real-world applications. Unfortunately, they may show "Clever Hans"-like behavior—making use of confounding factors within datasets—to achieve high performance. In this work, we introduce the novel learning setting of "explanatory interactive learning" (XIL) and illustrate its benefits on a plant phenotyping research task. XIL adds the scientist into the training loop such that she interactively revises the original model via providing feedback on its explanations. Our experimental results demonstrate that XIL can help avoiding Clever Hans moments in machine learning and encourages (or discourages, if appropriate) trust into the underlying model.

Previous plant phenotyping studies have focused on the visible (VIS, 400–700 nm), near-infrared (NIR, 700–1000 nm) and short-wave infrared (SWIR, 1000–2500 nm) range. The ultraviolet range (UV, 200–380 nm) has not yet been used in plant phenotyping even though a number of plant molecules like flavones and phenol feature absorption maxima in this range. In this study an imaging UV line scanner in the range of 250–430 nm is introduced to investigate crop plants for plant phenotyping. Observing plants in the UV-range can provide information about important changes of plant substances. To record reliable and reproducible time series results, measurement conditions were defined that exclude phototoxic effects of UV-illumination in the plant tissue. The measurement quality of the UV-camera has been assessed by comparing it to a non-imaging UV-spectrometer by measuring six different plant-based substances. Given the findings of these preliminary studies, an experiment has been defined and performed monitoring the stress response of barley leaves to salt stress. The aim was to visualize the effects of abiotic stress within the UV-range to provide new insights into the stress response of plants. Our study demonstrated the first use of a hyperspectral sensor in the UV-range for stress detection in plant phenotyping.

Interactions between different barley genotypes and the fungal pathogen Blumeria graminis f.sp. hordei (Bgh) have a specific impact on the crop physiology. Within the context of plant resistance phenotyping, it is relevant to investigate early host–pathogen interactions to avoid the crop infestation. Analyzing different parameters of the photosynthesis apparatus gives in-depth information of the plant’s health status and can be used for a spatial and temporal assessment of interaction types during plant–pathogen infestation. In the present study, experiments were performed with a near-isogenic line of barley cv. Ingrid WT (susceptible), mlo3 (papilla-based resistance) as well as a near-isogenic line of cv. Pallas, containing the Mla1 (hypersensitive response-based resistance) gene. After inoculation with Bgh isolate K1, the leaves were measured daily using chlorophyll fluorescence imaging. Inoculated, susceptible wild-type leaves showed a reduced effective quantum yield of the photosystem II (ΦPSII) already 1 day after inoculation. In accordance with the quantum yield reduction, the non-photochemical quenching (NPQ) increased, indicating thermal dissipation of excess energy. The changes of ΦPSII and NPQ represent modifications of the leaf metabolism to aid the fungal nutrition uptake, which is influenced by Bgh. By analyzing these parameters, it was also possible to indicate resistance reactions of mlo3 and Mla1 barley genotypes against Bgh. During papilla formation in mlo3 leaves, ΦPSII revealed the lowest values. In contrast, inoculated Mla1 leaves showed the lowest NPQ. The present study proofs that chlorophyll fluorescence imaging is a valuable tool for investigating early plant–pathogen interaction noninvasively. Furthermore, this phytopathology study uses chlorophyll fluorescence imaging, chlorophyll extraction and microscopic observations to characterize the interaction response of different genotypes to an Bgh infection.

Differences in early plant–pathogen interactions are mainly characterized by using destructive methods. Optical sensors are advanced techniques for phenotyping host–pathogen interactions on different scales and for detecting subtle plant resistance responses against pathogens. A microscope with a hyperspectral camera was used to study interactions between Blumeria graminis f. sp. hordei and barley (Hordeum vulgare) genotypes with high susceptibility or resistance due to hypersensitive response (HR) and papilla formation. Qualitative and quantitative assessment of pathogen development was used to explain changes in hyperspectral signatures. Within 48 h after inoculation, genotype-specific changes in the green and red range (500 to 690 nm) and a blue shift of the red-edge inflection point were observed. Manual analysis indicated resistance-specific reflectance patterns from 1 to 3 days after inoculation. These changes could be linked to host plant modifications depending on individual host–pathogen interactions. Retrospective analysis of hyperspectral images revealed spectral characteristics of HR against B. graminis f. sp. hordei. For early HR detection, an advanced data mining approach localized HR spots before they became visible on the RGB images derived from hyperspectral imaging. The link among processes during pathogenesis and host resistance to changes in hyperspectral signatures provide evidence that sensor-based phenotyping is suitable to advance time-consuming and cost-expensive visual rating of plant disease resistances.

Differences in early plant-pathogen interactions are mainly characterized by using destructive methods. Optical sensors are advanced techniques for phenotyping host-pathogen interactions on different scales and for detecting subtle plant resistance responses against pathogens. A microscope with a hyperspectral camera was used to study interactions between Blumeria graminis f. sp. hordei and barley (Hordeum vulgare) genotypes with high susceptibility or resistance due to hypersensitive response (HR) and papilla formation. Qualitative and quantitative assessment of pathogen development was used to explain changes in hyperspectral signatures. Within 48 h after inoculation, genotype-specific changes in the green and red range (500 to 690 nm) and a blue shift of the red-edge inflection point were observed. Manual analysis indicated resistance-specific reflectance patterns from 1 to 3 days after inoculation. These changes could be linked to host plant modifications depending on individual host-pathogen interactions. Retrospective analysis of hyperspectral images revealed spectral characteristics of HR against B. graminis f. sp. hordei. For early HR detection, an advanced data mining approach localized HR spots before they became visible on the RGB images derived from hyperspectral imaging. The link among processes during pathogenesis and host resistance to changes in hyperspectral signatures provide evidence that sensor-based phenotyping is suitable to advance time-consuming and cost-expensive visual rating of plant disease resistances.

The detection and identification of plant diseases is crucial for an appropriate and targeted application of plant protection measures in crop production. Recently, intensive research has been conducted to develop innovative and technology-based optical methods for plant disease detection. In contrast to common visual rating and detection methods, optical sensors are able to measure pathogen-induced changes in the plant physiology non-invasively and objectively. Several studies showed that especially hyperspectral sensors are valuable tools for disease detection, identification and quantification on different scales from the tissue to the canopy level. This review describes the basic principles of hyperspectral measurements and different types of available hyperspectral sensors. Possible applications of hyperspectral sensors on different scales for disease detection and plant protection are discussed and evaluated. The advantages and disadvantages on each particular scale, as well as the impact of external factors, such as: light, wind, viewing angle, for measurements in laboratories, greenhouses and fields, are critically assessed in order to support researchers and agriculture technicians. Additionally, a comprehensive literature review about the use of hyperspectral sensors on these different scales for plant disease detection reflects the possibilities of non-invasive measurement systems. This highlights advantages of hyperspectral sensors when investigating plant–pathogen interactions through multiple examples. By some approaches, detection before visible symptoms appear is feasible. The potential of hyperspectral sensors as a tool for disease identification and quantification, based on disease characteristic changes in the plants spectral signature, is discussed as well. The review is concluded with an overview on different data analysis methods, which are required to extract key information from gathered hyperspectral datasets.