The worldwide food demand is expected to increase by 35 % to 56 % between 2010 and 2050 due to the growing world population. At the same time, about a third of earth’s land surface is already used for agriculture. To increase agricultural production, without expanding the land under cultivation or increasing agrochemical inputs used, sustainable ways of intensification are necessary. One approach of sustainable intensification is to breed high-performance genotypes that are well adapted to specific environments, and breeding was instrumental in the sharp increase in yields since the Green Revolution. Nevertheless, yields of important field crops, including wheat (Triticum aestivum L.), have stagnated since the late 1990s. This has significant implications for future food security, as wheat is one of the most important staple crops. Calories from wheat supplied up to 21 %of the energy consumed by humans. Genetic gain of wheat was shown to not have declined, and under optimal conditions, increased grain yields can still be realized. But genetic gains were in part counteracted by climate change, which comes with a higher frequency of adverse growing conditions such as drought and heat during sensitive growing stages and leads to a climate that is generally less favorable for agriculture, especially in temperate and hot climates. Thus, further efforts should focus on adapting genotypes and management practices to local conditions and climates, as interactions between genotypes and environments (GxE) but also management (GxExM) are responsible for large variability in grain yields. Developing and identifying optimal genotypes for specific environments is paramount to closing the gap between the attainable and the realized yield. Plant breeding allows for the development of improved varieties. To translate genetic progress in breeding into higher yields, the most suitable genotypes must be used in specific environments. Thus, breeding must be paralleled with a thorough characterization of variety performance in respective environments by conducting multi-environment trials (MET) for variety testing. Results of variety testing are published in annual lists of recommended varieties to allow farmers and other stakeholders to choose varieties that meet the market goals in their specific environments and soils. Typical traits monitored in variety testing are grain yield at 15 % water content, lodging resistance, early maturity, early heading, sprouting, overwintering, plant height, thousand kernel weight, hectoliter weight, resistance to various diseases such as powdery mildew, rusts, different Septoria species and Fusarium head blight. On harvested grains, the baking quality, the sedimentation index (Zeleny test), and the protein content are evaluated. These traits are typically still assessed by field observations or laboratory analysis, which is labor intensive and costly, especially since variety testing usually uses METs and thus traits must be assessed on multiple sites. In breeding, high-throughput field phenotyping (HTFP) methods were proposed and developed to make the assessment of plant traits more efficient and also to assess novel traits. Variety testing could also profit from new HTFP methods, but many of these methods have been tested experimentally under relatively controlled conditions and still have a relatively low technology readiness level (TRL). They have thus not yet been established in the daily practice or variety testing. To close the gap between basic research and methods that are actually applied by variety testing organizations to finally benefit farmers’ production, translational research is necessary. This thesis focuses on “lean phenotyping” as one aspect of translational research in the context of variety testing. Many of the proposed HTFP workflows are just too expensive in terms of initial investments, operational costs, and labor to be applied in variety testing, especially as within MET, multiple sites must be measured. Switching to cheaper equipment or simplifying workflows is often not easily possible, as the quality of the measured traits is too poor for beneficial integration into variety testing. Lean phenotyping in this context is understood as translational research to design workflows in such a way that results of sufficient quality can be generated even with more affordable sensors or that the costs of high-quality methods can be reduced. Another aspect concerns the use of new technologies or sensors to measure new traits, provided that they add value for variety testing. The ultimate goal in lean phenotyping is to develop methods with an acceptable balance of costs and benefits. This thesis was carried out within the Agroscope “Production Technology & Cropping Systems Group”, which is responsible for the official variety testing of wheat and other cereals in Switzerland. The thesis is committed to translational research on phenotyping methods for variety testing to further develop them toward lean phenotyping. It aims to evaluate and increase the TRL of existing phenotyping methods under realistic variety testing conditions. Therefore, three optical lean phenotyping methods, drone-based thermal cameras, PhenoCams, and chlorophyll fluorescence sensors were developed, adapted, and examined. Airborne thermography is a promising method for measuring canopy temperature (CT) to examine the relative fitness of a plant in an environment, especially in the context of heat and drought. With the development of drone-based thermal cameras, airborne thermography became easily accessible and affordable. However, the high variability of CT data from such uncooled thermal cameras makes interpretation very challenging and hindered the broad adoption of this new technology. Therefore, a multi-view approach was adapted for drone-based thermal cameras. Without changing the equipment used, but only with a novel and more comprehensive statistical analysis pipeline, the temporal and spatial variability of CT could be estimated and corrected, allowing for more genotype-specific and consistent measurements. This increased the interpretability of CT, thereby rendering thermal imaging more applicable and therefore more interesting as a phenotyping method in wheat variety testing. However, CT is an ephemeral trait and influenced by many factors in the short term. The thermal sensor and CT itself are very sensitive to confounding environmental influences. In addition, viewing geometry related effects add uncertainty to CT estimates. These effects mask experimental sources of variance, such as different genotypes and treatments, and while CT is mostly considered a proxy measure of stomatal conductance, the trait also features phenotypic correlations with other traits such as plant height or fractional canopy cover. To gain a thorough understanding of CT as a trait, the sources of variance of drone-based uncooled thermography were thoroughly examined based on 99 flights. Using the thermal multi-view approach developed in the previous step, more than 96.5 % of the initial variance could be explained on average by experimental and confounding sources of variance combined. The insights gained support the planning, conducting, and interpretation of drone-based CT screenings in variety testing. While drone-based CT represents a new trait that could be useful in the context of evaluating varieties and their resistance to drought and heat, this thesis also developed methods to screen established traits more efficiently and objectively. It is crucial to know the timing of phenological stages and the senescence behavior of genotypes to select for locally adapted varieties and to plan crop management accordingly. Knowing the timing of phenological stages also allows for a more meaningful interpretation of G⇥E interactions. Capturing these traits with frequent field visits is very time-consuming. A semimobile PhenoCam setup was used to track phenology and senescence from ear emergence to full maturity. An economic analysis revealed that PhenoCams are economically interesting for observing distant experimental sites. Thus, PhenoCams offer a cost-effective replacement of visual ratings of phenology and senescence, especially in the context of MET. As for evaluating the timing of phenological stages and senescence, the rating of plant disease infestations under field conditions is time-consuming and prone to subjectivity. Chlorophyll fluorescence (CF) was proposed as a tool to track and rate Fusarium head blight infestations and this chapter explored the potential and limitations of CF methods under field conditions. A hand-held CF sensor was used to track Fusarium infestations first in a greenhouse trial and the method was then transferred to a field trial and tested for two seasons, together with a CF imaging approach. The tested methods worked well in high-level infestations, but it is hypothesized that they would fail at low-level infestations due to a too low number of measurements and the throughput of the method would need to be increased drastically, e.g. by automatization with field robots. This work provides methodologies and insight for three optical lean phenotyping methods in the context of wheat variety testing. For drone-based thermography, a novel statistical approach was developed to handle the large variability of such data and the approach was applied to examine the manifold sources of variance in CT estimates based on thermal images. PhenoCams were applied to observe phenology and senescence and finally the potential of chlorophyll fluorescence to track the disease progression of Fusarium in field conditions was examined.