Achieving global goals for sustainable nutrition, health, and wellbeing will depend on delivering enhanced diets to humankind. This will require instantaneous access to information on food-source quality at key points of agri-food systems. Although laboratory analysis and benchtop NIR spectrometers are regularly used to quantify grain quality, these do not suit all end users, for example, stakeholders in decentralized agri-food chains that are typical in emerging economies. Therefore, we explored benchtop and portable NIR instruments, and the methods that might aid these particular end uses. For this purpose, we generated NIR spectra for 328 grain samples from multiple cereals (finger millet, foxtail millet, maize, pearl millet, and sorghum) with a standard benchtop NIR spectrometer (DS2500, FOSS) and a novel portable NIR-based instrument (HL-EVT5, Hone). We explored classical deterministic methods (via winISI, FOSS), novel machine learning (ML)-driven methods (via Hone Create, Hone), and a convolutional neural network (CNN)-based method for building the calibrations to predict grain protein out of the NIR spectra. All of the tested methods enabled us to build relevant calibrations out of both types of spectra (i.e., R² ≥ 0.90, RMSE ≤ 0.91, RPD ≥ 3.08). Generally, the calibration methods integrating the ML techniques tended to enhance the prediction capacity of the model. We also documented that the prediction of grain protein content based on the NIR spectra generated using the novel portable instrument (HL-EVT5, Hone) was highly relevant for quantitative protein predictions (R² = 0.91, RMSE = 0.97, RPD = 3.48). Thus, the presented findings lay the foundations for the expanded use of NIR spectroscopy in agricultural research, development, and trade.

Finger millet is an important drought-tolerant and grain-nutrient dense food crop grown in semi-arid regions in Asia and Africa. The forage is used as a source of dry roughage for feeding livestock. In this study, the finger millet diversity panel (310 accessions and four controls) representing the global collection of the finger millet germplasm conserved at the ICRISAT genebank was assessed for forage quality and diversity in the years 2018 and 2019. Results of the study suggested that finger millet can generate stover yield ranging from 2890 to 10,779 kg ha⁻¹. Finger millet forage contained 6.47% to 8.15% of crude protein, >90% of dry matter content, 11.47% to 14.17% of ash content, 62.36% to 67.73% of neutral detergent fiber (NDF), 33.07% to 38.37% of acid detergent fiber (ADF), 3.95% to 4.80% of acid detergent lignin (ADL), 6.18% to 6.89% of metabolizable energy (ME) and 45.21% to 49.09% of in vitro organic matter digestibility (IVOMD) with the grain yield of 810 to 3698 kg ha⁻¹ at maturity stage. The per se performance between the races, regions, and biological status has been performed and differed significantly for important traits. The 314 accessions were grouped into five clusters based on the performance for food-feed traits. Except for crude protein, there was a positive relationship between forage quality-positive traits and grain yield, indicating that agronomic and forage quality traits could be improved simultaneously. The top 10 promising accessions for important forage quality traits and accessions with multiple forage quality traits were identified. This study provides a detailed understanding of the variability that exists in forage quality traits in crop residues and their association with agronomic traits in the finger millet germplasm. The identified top-performing accessions would be the key genetic resources for developing dual-purpose cultivars and the information from this study will be useful for future finger millet food-feed trait improvement.

Limited supply of quality feed is the most common problem limiting livestock productivity in sub‐Saharan Africa (SSA). Routine feed evaluation is indispensable for formulating balanced rations, feed characterization, safety, and minimizing the environmental impact of livestock. Traditional wet chemistry has not met this demand in SSA because it is time consuming, expensive, reliant on imported reagents and equipment that requires regular maintenance. Near infrared reflectance spectroscopy (NIRS) is a rapid and accurate alternative. The NIRS can help meet the need to characterize locally available forages and feeds on the continent, thus allowing formulation of optimally balanced and safe rations, facilitating establishment of nutritive value‐based pricing, and improving feed marketing and environmental stewardship. Though several NIRS systems have been purchased in many SSA countries, few are currently used. Reasons include high upfront costs, lack of requisite technical capacity, lack of access to comprehensive wet chemistry‐based databases to develop and validate robust and accurate predictive equations, lack of access to or relevance of existing validated equations, and limited awareness about the value of NIRS. Recently developed portable devices can dramatically reduce cost, while providing flexibility and comparable accuracy to benchtop systems. Formation of NIRS consortia and communities of practice including public–private partnerships that link equipment, pool resources, and provide periodic training and troubleshooting, can address many of these problems. This paper elaborates the potential for using NIRS to improve feed analysis in SSA countries, the reasons for the low use of existing systems, and strategies to improve the adoption and use of NIRS.

Non-invasive crop analysis through image-based methods holds great promise for applications in plant research, yet accurate and robust trait inference from images remains a critical challenge. Our study investigates the potential of AI model ensembling and hybridization approaches to infer sorghum crop traits from RGB images generated via unmanned aerial vehicle (UAV). In our study, we cultivated 21 sorghum cultivars in two independent seasons (2021 and 2022) with a gradient of fertilizer and water inputs. We collected 470 ground-truth N measurements and captured corresponding RGB images with a drone-mounted camera. We computed five RGB vegetation indices, employed several ML models such as MLR, MLP, and various CNN architectures (season 2021), and compared their prediction accuracy for N-inference on the independent test set (season 2022). We assessed strategies that leveraged both deep and handcrafted features, namely hybridized and ensembled AI architectures. Our approach considered two different datasets collected during the two seasons (2021 and 2022), with the training set from the first season only. This allowed for testing of the models’ robustness, particularly their sensitivity to concept drifts, in the independent season (2022), which is fundamental for practical agriculture applications. Our findings underscore the superiority of hybrid and ensembled AI algorithms in these experiments. The MLP + CNN-VGG16 combination achieved the best accuracy (R² = 0.733, MAE = 0.264 N% on an independent dataset). This study emphasized that carefully crafted AI-based models applied to RGB images can achieve robust trait prediction with accuracies comparable to the similar phenotyping tasks using more complex (multi- and hyper-spectral) sensors presented in the current literature.

Despite a long history promoting improved forages, their adoption in sub‐Saharan Africa remains limited. Identifying cultivars and developing the right agronomic practices suited to specific forage systems can help reduce the feed gap in this region. This 2‐yr study was performed under rainfed conditions at the Institute of Environment and Agricultural Research research station in Farakoba, Burkina Faso. Treatments included two planting methods (continuous rows 40‐cm apart vs. open pits spaced 40 × 40 cm) and three perennial grasses: Urochloa spp. cultivar Mulato II, Congo grass [U. ruziziensis (R. Germ. and C.M. Evrard) Crins] and Guinea grass [Megathyrsus maximus (Jacq.) B.K. Simon & S.W.L. Jacobs] cultivar C1. The experimental design was a randomized complete block under a split plot treatment arrangement, and planting method was the main plot (3 × 7 m) and the grass varieties the subplots (3 × 3 m). Plots were established in August 2018, and herbage accumulation (HA) and nutritive value (NV) were measured in 2018 and 2019. Regardless of planting method, Mulato II had greater total herbage mass (26 Mg ha^–1) after 2 yr than did Congo grass (15 Mg ha^–1) and cultivar C1 (12 Mg ha^–1) (p ≤ .05). Mulato II had the greatest crude protein (170 g kg^–1), in vitro digestible organic matter (635 g kg^–1), and metabolizable energy (89 MJ kg^–1), and least neutral detergent fiber (537 g kg^–1) concentrations, as compared to other cultivars (p ≤ .05). Both HA and NV were greater (p ≤ .001) for open pits spaced 40‐cm apart than the continuous rows for all cultivars, indicating that this planting method seems more appropriate to establish forage grasses. Mulato II showed potential to increase forage productivity and NV in Burkina Faso.

Twenty six lines each of two legume crop species: mungbean (Vigna radiata (L.) R. Wilczek) and vegetable soybean (Glycine max (L.) Merr.) were evaluated for grain yield and fodder traits during 2013 in Hyderabad, India. Three best performing lines and one poor performing line were selected in each species for further testing during 2014. Line x year interaction was significant for neutral detergent fibre, acid detergent fibre, acid detergent lignin, metabolisable energy and in vitro organic matter digestibility for mungbean. In the case of soybean, line x year interaction was significant for haulm yield, seed yield, nitrogen content, metabolisable energy and in vitro organic matter digestibility. Mungbean lines, VC 6510-151-1 and ML818 were the best for seed yield and fodder traits. Vegetable soybean lines, Swarna Vasundhra and AGS 406 were the best performers for fresh pod yield, haulm yield and in vitro organic matter digestibility. In-vivo parameters such as dry (DMI) and organic matter intake (OMI), dry (DMD) and organic matter digestibility (OMD) and nitrogen (N) balance were determined after feeding sheep with haulms from selected lines of mungbean and vegetable soybean as supplements to a sorghum stover based ration. Vegetable soybean line, Swarna Vasundhra recorded the highest OMD while the mungbean line, VC 6510-151-1 recorded the highest nitrogen balance.