The soybean cyst nematode (SCN) is the most important soybean pathogen in the United States with 3 billion dollars annual losses, exceeding the total losses of the next seven soybean yield robbing pathogens combined. While there are management options available for this pathogen, the ever-adapting nature of SCN populations in production fields puts extensive pressure on available management tools, such as the use of soybean varieties resistant to SCN. Thus, the proper management of SCN requires the adoption of an integrated approach using a multitude of control measures for efficient and sustainable management of this pest. A critical need for sustainable and effective management of soilborne pathogens is to have a precise assessment of the inoculum potential in the soil. However, the methods that are currently used for nematode quantification are crude with many limitations, such as the lack of accuracy and efficiency at the field or farm level. Also, crop damage thresholds are difficult to establish for nematodes due to the influence of heterogeneous soil chemical and physical properties, the interaction with other pathogens, and the effects of environmental factors on the ability of the nematode to cause losses

Recent advances in multispectral imaging and modeling make remote sensing a promising and attractive alternative to traditional crop disease detection and monitoring methods owing to its flexibility, low costs, and effectiveness. Multispectral and hyperspectral remote sensing technologies have been widely used to detect crop diseases (Calderón et al., 2013; Di Gennaro et al., 2016; Hatton et al. 2017), when the leaf canopy exhibits unique spectral signatures of crop stresses in response to different diseases and infection levels. These technologies have been applied to a wide range of sensors and platforms, including handheld spectroradiometer systems, Unmanned Aerial Systems (UAS), aircrafts and satellites imaging systems, which allow the diseases to be detected in a wide range of scales. However, there is still a general paucity of research on the combined use of such tools to monitor SCN and other pathogens of soybean.

Spatial technologies such as remote sensing and geographic information systems (GIS)-based tools are crucial to optimize and streamline the decision-making process in crop production and are being increasingly integrated to more advanced pathogen management systems. Nutter et al. (2002) evaluated GIS and remote sensing technologies that can be used to detect SCN related damage and identify the extent and distribution of the pathogen across fields. Groundbased data was collected using a multispectral radiometer. Aerial images were collected by a RGB and infrared camera attached to a plane. Aerial and ground-based data was collected every ten to fourteen days during the growing season. Also, the data was supplemented by Landsat satellite imagery. The results show that the data collected by integrating three imaging platforms explained more variation in yield than the traditional testing methods for SCN populations in the targeted soybean-production fields. A more recent study, targeting the sugar beet cyst nematode (Heterodera schachtii), revealed that multispectral imagery collected with UAVs could differentiate the areas of high nematode infestation across a field (Joalland et al., 2018). The study also found that the canopy temperatures of susceptible sugar beet cultivars tend to be higher compared to those recorded with tolerant cultivars. Differences in cultivars were also observed for biomass, chlorophyll content, and general stress. These differences were identified using Normalized Difference Vegetation Index (NDVI), chlorophyll content (CHLG), and healthy-index (HI) indices developed from UAV images and field spectroradiometer measurements. Note that the sugar beet cyst nematode has a similar biology and parasitism strategy to the soybean cyst nematode. We propose to expand on these studies by developing and optimizing plant health monitoring algorithms and associated predictive modeling tools to assess SCN infestation levels and their effect on plant health. These tools would be made available to the public to allow broad adoption for the benefit of Illinois soybean producers. The developed protocols are expected to be adapted to address other stresses caused by soilborne pathogens and pests.

1. To determine if UAV-based remote sensing tools can reliably predict and monitor nematode infestation levels and distribution in production fields

2. To develop usable AI algorithms and tools to assess the impacts of SCN and other soilborne pathogens on plant health

The proposal was developed based on preliminary research activities with promising results in the field. In 2019, we collected preliminary data from a 70 acre field in Carmi, IL in support of this project. A Matrice 210 UAV was used to collect imagery with a X5S RGB camera and an Altum Multispectral/Thermal camera. Yield data was collected, and the field was grid mapped and sampled on 1/3 acre grids resulting in 224 soil samples. Preliminary analysis reveal that SCN egg counts are correlated with colors and patterns in UAV images associated with greater plant stress. Figure 1 is a RGB image of a field in Carmi, one of the proposed locations. The areas highlighted in the image are suffering due to nematode feeding. Figure 2 shows the intensity of SCN infestation in this field ranging from a low of 1,300 eggs/100cc soil to over 40,000 eggs /100 cc soil. We are completing data analysis and refining our models and techniques with this test field for the proposed trials.

For this project, field trials will be established at two locations in southern Illinois. One location will be in Carmi, IL and second location will be identified that represents a different soil type.
These locations will serve as data generators for the predictive models. Two different experimental setups will be used in the field trials:

First experimental setup - The experiment will be conducted at two locations. Soybean will be planted in SCN infested fields. These fields will represent traditional fields in Illinois with varying infestation level of SCN.

Second experimental setup - The study will be conducted at two locations and will consist of 24 rows treated with a fungicide and ILEVO seed treatment and 24 rows with only a fungicide treatment. The two treatments (with nematicide and without) will be replicated 4 times.

Fields that have a range of low (0 - 500 eggs per 100 cc soil) to high (<15,000 eggs per 100 cc of soil) will be selected. The field sizes will range from 70 – 150 acres. The fields will be mapped into 1/3-acre grids, where soil samples will be collected. Soil samples will be collected at planting, midseason, and harvest to determine the population densities of SCN and for a soil fertility analysis. By determining the SCN population densities, we will be able to validate the areas in the field with high populations and relate those populations to the imagery. At the end of the season, soybeans will be harvested, and the yield data will be collected from yield monitors. Multispectral and visual UAV data will be collected at each field site, along with close-range spectral data collected with a spectroradiometer. Images will be collected using an X5S RGB camera and an Altum multispectral camera mounted to a UAV. In the strips where there are different treatments, 20 spectral readings will be collected using a spectroradiometer, and the data will then be averaged to represent each strip. Imagery and spectra will be collected 30, 45, 60, and 75 DAP (days after
planting).

Collection of multispectral data - Every flight mission will be designed with 80% forward overlaps and 80% side overlaps. All flight missions will be completed under the local drone related legislative and campus regulations. Each flight mission will cover the entirety of the trial area and the time of flight will be dependent on the size of the area. To ensure that the imagery will be correctly georeferenced, ground control points (GCPs) will be used to correctly project the imagery. The GCPs will be placed randomly throughout the study area and the GPS coordinates will be collected. RTK correction will be used to give sub-centimeter accuracy of the GCPs.

Development of AI algorithms - As shown in previous research that nematode (e.g., Heterodera schactii) activities present unique spectral and textual signatures on the phenological representation of soybean fields, we propose to develop agricultural artificial intelligence (Agro-AI) models to identify the areas with high nematode densities for better management of SCN. The Agro-AI models are especially suitable to identify these fuzzy-look features that are difficult to be examined with regular field or image inspection. This study will collect UAV image samples to develop deep learning models. The manually labeled UAV images containing both nematodeaffected and healthy areas will be randomly divided 2/3 and 1/3 as the training and testing samples respectively. Data augmentation techniques such as rotation, reflection, and Gaussian smoothing kernel filtering will be applied to the original training samples to increase the size of the training dataset and prevent the deep neural network from overfitting. The training samples will then be used to feed a deep learning convolutional neural network (CNN) algorithm. The Faster R-CNN detector can be trained by the layers with ten epochs. The haar-like cascading detector can be trained by 50 training stages with a 0.01 false alarming rate. Every testing sample is treated with the well-trained faster R-CNN detector and cascading detector. Their corresponding CNN feature and haar-like feature could be extracted. The likelihood of each feature reflecting nematodes infection are predicted by the detector. The non-maximum suppression (NMS) will be used to suppress redundant detections surrounding the same feature. Every training image is sliced to 9 equal size small images to enhance training efficiency. Final training samples used to train the model are enlarged to 5 times the original data. A valid detection is defined by an overlap ratio of the predicted bounding box and ground truth box that is beyond 0.4.

Testing samples will be labeled as three groups based on the severity of the nematodes’ destructive effects based on expert interpretation. Accuracy on every single image and corresponding
detection method will be recorded as the response and method factor, respectively. As the response is bounded from 0 to 1, a beta Generalized Linear Model (GLM) is fitted to interpret the treatment effect of the detection methods and environmental types on detection accuracy. The final precision and recall of the detection by every testing sample will be recorded. The F-score, a measure of a
test’s accuracy considering both the precision P and the recall R, is calculated to assess the performance of this model. The percentage of training samples varying from 10%-100% with 10%
interval can be used to determine the minimum training data size required for accurate detection.

Our main goal is to develop a UAS-based artificial intelligence (AI) toolkit to detect SCN related stresses of soybean. This project will focus on developing a UAS-based remote sensing predictive modeling tool to determine SCN population densities and their impact on soybean health. The tool would also be useful in assessing the efficiency of pesticides, resistant varieties cultural control practices and other new approaches to manage SCN and to identify the optimal conditions for their deployment. The development of this toolkit will increase our ability to manage SCN by being more precise in our deployment of management tools in production fields. This project is expected to advance the research on plant health monitoring algorithms and tools that can remotely assess SCN infestation levels and predict their effect on plant health and yield.

Through the collection of drone images and ground truth SCN data, this project will develop a set of algorithms and tools to assess the impact of SCN and potentially other soilborne pathogens on plant health. We expect that the developed algorithms and tools can be applied in Illinois and other Midwest states.