<p>High-throughput phenotyping using high spatial, spectral, and temporal resolution remote sensing (RS) data has become a critical part of the plant breeding chain focused on reducing the time and cost of the selection process for the “best” genotypes with respect to the trait(s) of interest. In this study, the potential of accurate and reliable sorghum biomass prediction using hyperspectral and LiDAR data acquired by sensors mounted on UAV platforms is investigated. Experiments comprised multiple varieties of grain and forage sorghum, including some photoperiod sensitive varieties, providing an opportunity to evaluate a wide range of genotypes and phenotypes. </p><p>Feature extraction is investigated, where various novel features, as well as traditional features, are extracted directly from the hyperspectral imagery and LiDAR point cloud data and input to classical machine learning (ML) regression based models. Predictive models are developed for multiple experiments conducted during the 2017, 2018, and 2019 growing seasons at the Agronomy Center for Research and Education (ACRE) at Purdue University. The impact of the regression method, data source, timing of RS and field-based biomass reference data acquisition, and number of samples on the prediction results are investigated. R2 values for end-of-season biomass ranged from 0.64 to 0.89 for different experiments when features from all the data sources were included. Using geometric based features derived from the LiDAR point cloud and the chemistry-based features extracted from hyperspectral data provided the most accurate predictions. The analysis of variance (ANOVA) of the accuracies of the predictive models showed that both the data source and regression method are important factors for a reliable prediction; however, the data source was more important with 69% significance, versus 28% significance for the regression method. The characteristics of the experiments, including the number of samples and the type of sorghum genotypes in the experiment also impacted prediction accuracy. </p><p>Including the genomic information and weather data in the “multi-year” predictive models is also investigated for prediction of the end of season biomass. Models based on one and two years of data are used to predict the biomass yield for the future years. The results show the high potential of the models for biomass and biomass rank predictions. While models developed using one year of data are able to predict biomass rank, using two years of data resulted in more accurate models, especially when RS data, which encode the environmental variation, are included. Also, the possibility of developing predictive models using the RS data collected until mid-season, rather than the full season, is investigated. The results show that using the RS data until 60 days after sowing (DAS) in the models can predict the rank of biomass with R2 values of around 0.65-0.70. This not only reduces the time required for phenotyping by avoiding the manual sampling process, but also decreases the time and the cost of the RS data collections and the associated challenges of time-consuming processing and analysis of large data sets, and particularly for hyperspectral imaging data.</p><p>In addition to extracting features from the hyperspectral and LiDAR data and developing classical ML based predictive models, supervised and unsupervised feature learning based on fully connected, convolutional, and recurrent neural networks is also investigated. For hyperspectral data, supervised feature extraction provides more accurate predictions, while the features extracted from LiDAR data in an unsupervised training yield more accurate prediction. </p><p>Predictive models based on Recurrent Neural Networks (RNNs) are designed and implemented to accommodate high dimensional, multi-modal, multi-temporal data. RS data and weather data are incorporated in the RNN models. Results from multiple experiments focused on high throughput phenotyping of sorghum for biomass predictions are provided and evaluated. Using proposed RNNs for training on one experiment and predicting biomass for other experiments with different types of sorghum varieties illustrates the potential of the network for biomass prediction, and the challenges relative to small sample sizes, including weather and sensitivity to the associated ground reference information.</p>

<p>High-throughput phenotyping has emerged as a powerful approach to expedite crop breeding programs. Modern remote sensing systems, including manned aircraft, unmanned aerial vehicles (UAVs), and terrestrial platforms equipped with multiple sensors, such as RGB cameras, multispectral, hyperspectral, and infrared thermal sensors, as well as light detection and ranging (LiDAR) scanners are now widely used technologies in advancing high throughput phenotyping. These systems can collect high spatial, spectral, and temporal resolution data on various phenotypic traits, such as plant height, canopy cover, and leaf area. Enhancing the capability of utilizing such remote sensing data for automated phenotyping is crucial in advancing crop breeding. This dissertation focuses on developing deep learning and transfer learning methodologies for crop phenotyping using multi-modality remote sensing and environmental data. The techniques address two main areas: multi-temporal/across-field biomass prediction and multi-scale remote sensing data fusion.</p> <p><br></p> <p>Biomass is a plant characteristic that strongly correlates with biofuel production, but is also influenced by genetic and environmental factors. Previous studies have shown that deep learning-based models are effective in predicting end-of-season biomass for a single year and field. This dissertation includes development of transfer learning methodologies for multiyear,</p> <p>across-field biomass prediction. Feature importance analysis was performed to identify and remove redundant features. The proposed model can incorporate high-dimensional genetic marker data, along with other features representing phenotypic information, environmental conditions, or management practices. It can also predict end-of-season biomass using mid-season remote sensing and environmental data to provide early rankings. The framework was evaluated using experimental trials conducted from 2017 to 2021 at the Agronomy Center for Research and Education (ACRE) at Purdue University. The proposed transfer learning techniques effectively selected the most informative training samples in the target domain, resulting in significant improvements in end-of-season yield prediction and ranking. Furthermore, the importance of input remote sensing features was assessed at different growth stages.</p> <p><br></p> <p>Remote sensing technology enables multi-scale, multi-temporal data acquisition. However, to fully exploit the potential of the acquired data, data fusion techniques that leverage the strengths of different sensors and platforms are necessary. In this dissertation, a generative adversarial network (GAN) based multiscale RGB-guided model and domain adaptation framework were developed to enhance the spatial resolution of multispectral images. The model was trained on limited high spatial resolution images from a wheel-based platform and then applied to low spatial resolution images acquired by UAV and airborne platforms.</p> <p>The strategy was tested in two distinct scenarios, sorghum plant breeding, and urban areas, to evaluate its effectiveness.</p>

<p>Cereal rye (<i>Secale cereale</i>, L., CR) is the most commonly utilized cover crop species within the United States. Yet, the total land area planted to CR on an annual basis remains relatively low despite its numerous proven environmental benefits. The relatively low rates of CR adoption could be due to a dearth of knowledge surrounding certain agronomic and economic components of CR adoption. Currently, there exists knowledge gaps within the scientific literature regarding CR performance, N cycling, and associated economic risk. <a>Thus, to address the above-mentioned knowledge gaps, three individual studies were developed to: i) investigate the fate of scavenged CR nitrogen (N) amongst soil N pools, ii) assess the suitability of visible-spectrum vegetation indices (VIs) to predict CR biomass and nutrient accumulation (BiNA), and iii) characterize the economic risk of CR adoption at a regional scale over time.</a></p> <p>In the first study, ¹⁵N, a stable isotope of N, was used in an aerobic incubation to track the fate of CR root and shoot N among the soil microbial biomass, inorganic, and organic N pools, as well as explore CR N bioavailability over a simulated corn growing season. In this study, the C:N ratio of the shoot residues was 16:1 and the roots was 31:1 and differences in residue quality affected the dynamics of CR N release from each residue type. On average, 14% of whole plant CR N was recovered in the soil inorganic N pool at the final sample date. Correspondingly, at the final sampling date 53%, 33%, and less than 1% of whole plant CR N was recovered as soil organic N, undecomposed residue, and as microbial biomass N, respectively. Most CR N remained unavailable to plants during the first cash crop growing season subsequent to termination. This knowledge could support the advancement of N fertilizer management strategies for cropping systems containing cereal rye.</p> <p>In the second study, a commercially available unmanned aerial vehicle (UAV) outfitted with a standard RGB sensor was used to collect aerial imagery of growing CR from which visible-spectrum VIs were computed. Computed VIs were then coupled with weather and geographic data using linear multiple regression to produce prediction models for CR biomass, carbon (C), N, phosphorus (P), potassium (K), and sulfur (S). Five visible-spectrum VIs (Visible Atmospherically Resistant Index (VARI), Green Leaf Index (GLI), Modified Green Red Vegetation Index (MGRVI), Red Green Blue Vegetation Index (RGBVI), and Excess of Green (ExG)) were evaluated and the results determined that MGRVI was the best predictor for CR biomass, C, K, and S and that RGBVI was the best predictor for CR N and P. Furthermore, the final prediction models for the VIs selected as the best predictors developed in this study performed satisfactorily in the prediction of CR biomass, C, N, P, K, and S producing adjusted R² values of 0.79, 0.79, 0.75, 0.81, 0.81, and 0.78, respectively. The results of this study have the potential to aid producers in making informed decisions regarding CR and fertility management. </p> <p>In the final study, agronomic data for corn and soybean cropping systems with and without CR was collected from six states (Illinois, Indiana, Iowa, Minnesota, Missouri, and Wisconsin) and used within a Monte-Carlo stochastic simulation to characterize the economic risk of adopting CR at a regional scale over time. The results of this study indicate that average net returns to CR are always negative regardless of CR tenure primarily due to added costs and increased variability in cash crop grain yields associated with CR adoption. Further, the results demonstrate that the additional risk assumed by adopting CR is not adequately compensated for with current CR adoption incentive programs and that the risk premium necessary can be 1.7 to 15 times greater than existing incentive payments. Knowledge gained from this study could be used to reimagine current incentive programs to further promote adoption of CR.</p>