<p> Significant portions of nitrogen (40–60%), phosphorus (80–90%) and potash (30–50%) applied in agricultural fields are not taken up by plants, causing serious issues for farmers and the environment. Fertilizer losses result in greater fertilizer input costs and the cost of fertilizer is projected to increase due to limited ore resources and increasing fossil fuel prices. Moreover, excess fertilizer application can contaminate water and air, resulting in human health problems. Leaching fertilizers also induce eutrophication, acid rain and global climate change. Therefore, developing crops with high nutrient uptake efficiency is important for economic and environmental sustainability of agriculture. Crop improvement depends on efficiency and accuracy of genotyping and phenotyping. Genotyping has improved in recent years and is generally efficient and accurate. In contrast, improvements in phenotyping lag far behind. Lack of high-throughput (efficient, accurate and inexpensive) phenotyping (HTP) methods limit the speed of genetic improvement. As a result, there is an increasing interest in development of HTP for predicting crop nutrient status. My research addresses whether hyperspectral data in the visible-near-infrared range (HS-VNIR) acquired by a handheld device or an unmanned aerial vehicle (UAV) can be used for predicting maize nutrient status. Proximal and remote sensing data coupled with ground reference measurements of hybrid maize nutrient status were collected in fertilizer strip trials conducted at Purdue Agricultural Centers located throughout Indiana. Statistical models were developed to predict nutrient status based on HS-VNIR with coefficients of determination of cross-validation [R² (CV)] used to evaluate the performance of the predictive models. Models with acceptable goodness-of-fit [R² (CV) > 0.30] were considered satisfactory. These studies demonstrated that models developed using handheld proximal sensing data performed adequately for predicting N, K, Mg, Ca, P, S, Mn, Zn and B. Similarly, models developed using UAV-based HS-VNIR could be used to predict N, K, Mg, Ca, P, S, Mn, Zn and B. Models that combine proximal and remote sensing data also performed well with predictions of N, K, Mg, Ca, P, S, Mn, Zn and B. In conclusion, handheld or UAV-based hyperspectral imaging can provide corn breeders with HTP data on the status of all macronutrients (N, K, Mg, Ca, P, S) and some micronutrients (Mn, Zn, B). Deployment of this technology may provide a valuable tool to support development of cultivars with improved nutrient uptake efficiencies.  </p>

This study aimed to develop a novel end-to-end plant diagnosis model for the analysis of plant health conditions in near real-time to optimize the rate of production on farmlands for an intensive, yet environmentally safe farming production to preserve the natural environment. First, field research was conducted to determine the extent of the problems faced by farmers in agricultural production. This allowed us to refine the research statement and the level of technology involved in the production processes. The advantages of unmanned aerial systems were exploited in the continuous monitoring of farm plantations to develop automated and accurate measures of farm conditions. To this end, this thesis applies the Precision Agricultural technology as a data based management system that takes into account spatial variations by using the Global Positioning System, Geographical Information System, remote sensing, yield monitors, mapping, and guidance system for variable rate applications. An unmanned aerial vehicle embedded with an optic and radiometric sensor was used to obtain high spectral resolution images of plantation status during normal production/growth cycle. Then, an ensemble of classifiers with Convolution Neural Networks (CNN) was used as off the shelf feature extractor to train images to develop an end-to-end feature detection and multiclass classification system for plant overall health’s conditions. Whereby previous works have concentrated on using CNN as off the shelf feature extractor and model training to detect only plant diseases from plants. To date, no research has yet been carried out to develop an end-to-end model for the overall plant diagnosis system. Previous studies focused on the detection of diseases at any given time, making it difficult to implement comprehensive real-time PA systems. Applying the pretrained model to the new images showed that the model can accurately predict any plant condition with an average of 97% accuracy.