Location: Crop Production Systems Research
2018 Annual Report
Objectives
Objective 1: Develop analytical methods and integrate them into decision support tools for effective aerial application.
Sub-objective 1.1: Verify new drift modeling paradigms with field data; optimize spray delivery systems for drift reduction considering temporal weather differences in statistical analysis.
Sub-objective 1.2: Determine plant injury due to off-target drift by spray sampling, biological measurements, and remote sensing.
Sub-objective 1.3: Determine periods of stable atmosphere favorable for long-distance movement of spray deleterious to susceptible crops downwind from spray application; quantify the effect of surface conditions with weather and incorporate this information into new guidelines for pilots to reduce potential for off-target movement of spray.
Objective 2: Develop laboratory, ground application, and aerial systems for delivery of biological control agents such as non-toxigenic A. flavus for control of mycotoxin and evaluate their effectiveness with bio-assay analysis.
Sub-objective 2.1: Determine field conditions that promote fungal contamination using on-the-go soil sensors and remote sensing; map risk zones for targeted application.
Sub-objective 2.2: Develop aerial application systems to deliver biological control agents and evaluate effectiveness of control with bio-assay analysis.
Objective 3: Develop methodologies that utilize existing remote sensing technologies for user- accessible agricultural aircraft and Unmanned Aerial System (UAS) platforms to detect invasive weeds and wild host plants for insect pests and distinguish between herbicide resistant and non-resistant weeds for use in selective spray management strategies.
Sub-objective 3.1: Identify spectral signatures and classification techniques to distinguish herbicide resistant from non-herbicide resistant weeds; evaluate imaging sensors using the identified signatures and map the distribution of herbicide resistant weeds for selective spraying.
Sub-objective 3.2: Identify spectral bands and classification techniques most useful in discriminating wild host plants of the tarnished plant bug from other land cover features and evaluate airborne imagery acquired to map these plants in and surrounding agricultural fields.
Sub-objective 3.3: Develop accessible remote sensing and rapid image processing systems for targeted application that can be operated by agricultural pilots; develop lightweight remote sensing systems requiring minimal user intervention for Unmanned Aerial Systems (UAS).
Approach
This project seeks to advance application technology through improvements in 1) drift management technologies and models; 2) aerial systems to effectively deliver biological control agents; and 3) remote sensing systems usable by pilots for agricultural aircraft to identify herbicide-damaged plants, invasive weeds, and wild host plants. While drift management is a concern for all pesticide applications, it is of particular concern for aerial applications. The use of herbicide-resistant crop varieties has increased use of glyphosate, both exacerbating the drift problem and giving rise to herbicide resistant weeds that need to be dealt with. Biological control is making headway, but aerial systems are needed to apply these agents. Aerial systems will be developed to effectively deliver liquid formulations of non-toxigenic biological agents to control mycotoxins in corn. Experiments for drift will attempt to reduce confounding of treatment data with environmental effects, preserving statistical precision of the experiments. Specific guidelines for pilots to prevent spraying during temperature inversions will be developed. The deleterious effects of off-target herbicide drift will be detected using spray and biological sampling, and hyperspectral and multispectral remote sensing. Remote sensing will also be used to detect herbicide resistant weeds and wild hosts for plant bug for targeted management. Improvements in remote sensing and rapid image analysis systems will allow accessibility of these systems by agricultural pilots. Autonomous Unmanned Aerial (or ”drone”) platforms will be developed with rapid image analysis capabilities for areas not served by agricultural aircraft. Experiments are also proposed to demonstrate the validity of techniques developed.
Progress Report
The study has been conducted to simulate spray deposit and downwind drift with the AgDisp spray drift model through the design of experiments. On the basis of simulation study field data on low-drift nozzles for different nozzle angles, orifice sizes, spray rates and release altitudes were further collected to validate the simulation. With these studies and analysis the protocols of testing and validation of the simulation models have been formulated.
The study has been conducted to assess soybean, cotton and corn injuries from off-target drift of aerially applied glyphosate by biological response measurements, spray sampling and aerial multispectral remote sensing. The studies were further extended to use of unmanned aerial vehicle (UAV) to detect non-dicamba tolerant soybean response to dicamba spray by low-altitude multispectral remote sensing combined with biological response measurement at different growth stages.
A website has been created and published for agricultural pilots in Mississippi Delta to have online advice the timing of spray to avoid spray drift caused by temperature inversions. The data used to calculate for the web site were collected from weather stations established in this areas.
Second year analyses of the imagery and corn ear data were completed for the corn aflatoxin study. For year three, airborne imagery was collected with the Tetracam multispectral camera system of study sites. Soil and vegetation zones were established with the data and computer software. Corn ears were sampled based on established soil and vegetation zones and sent to a laboratory for aflatoxin measurements.
Greenhouse and field studies were designed to compare glyphosate resistance in Palmer amaranth and Barnyard grass using hyperspectral plant sensing in continuing the previous studies on Palmer amaranth, Italian ryegrass and Johson grass. The studies were for rapid differentiation between glyphosate-resistant and glyphosate-sensitive weeds species through proximal hyperspectral remote sensing.
For the tarnished plant bug study, image acquisition with airborne system has been completed of study sites, and image analyses were completed with image processing software. Additionally, a study site was imaged with a UAV and data have been qualitatively accessed.
A new integrated imaging system with dual GoPro camera, FLIR thermal camera and Micasense RedEdge camera was developed and mounted to use on air tractor 402B and Micansense RedEdge-M, Parrot Sequoia and BaySpec hyperspectral cameras were mounted to use on different multirotor UAVs respectively to acquire hyperspectral, multispectral, thermal and color imagery to indicate crop vigor, canopy temperature and weed density and distribution areawide and over crop fields in the research farms.
Accomplishments
1. Web online service of temperature inversion determination for aerial applicators. ARS researchers in Stoneville, Mississippi, set up weather stations throughout the Stoneville, Mississippi, area. The data on wind speed, air temperature, and solar intensity measured at these weather stations were transferred wirelessly to a web site in the cloud. A web application was created from the data on the web site to provide online recommendations to aerial applicators in this area when the temperature inversion could occur to allow them to avoid the spray cloud flying out of the treated area. It is important for aerial applicators to work at the time to maximize the precision of spray targeting. The web application with the backend calculation from weather data provides a real-time, user-friendly access for aerial applicators to determine the atmospheric stability conditions at their locations in Mississippi Delta. This project has been continuously funded by Mississippi Soybean Promotion Board (MSPB) and the website has been published by MSPB to provide service for aerial applicators and producers.
2. Unmanned aerial vehicle (UAV) remote sensing of crop fields. UAV remote sensing significantly helps to improve crop field monitoring for precision agriculture with low cost, flexibility and high-resolution data. ARS researchers in Stoneville, Mississippi, have developed digital color, multispectral, hyperspectral and thermal imaging systems for being mounted on small UAVs. The applications of these systems have included crop injury from glyphosate/dicamba spray, crop plant height estimation, and crop yield estimation. Our UAV systems could cover any field on the research farm quickly and provide the images in a spatial resolution of a few centimeter. This year we continued the project that began last year to uniquely use UAVs to detect naturally-occurred glyphosate-resistant (GR) and glyphosate-susceptible (GS) weeds in soybean fields with digital color, multispectral and hyperspectral cameras at low altitude (20–30 m). The work is to transfer the previous greenhouse and field research results to regular crop fields to detect naturally-occurred GR and GS weeds to provide as-applied weed maps for site-specific weed management. This project is a collaboration with the Geosystems Research Institute, Mississippi State University and has been continuously funded by Mississippi Soybean Promotion Board (MSPB). Our research drew attention of domestic and international academia, industry and stakeholders and the ARS scientist has been invited to talk on the subject at international, national and regional meetings.
3. Mapping pigweed with free data and open source software. In the southeastern United States, pigweeds have become troublesome weeds in agricultural systems. To implement management strategies to control them, agriculturalists need information on areas affected by pigweeds. ARS scientists at Stoneville, Mississippi, used free and open-source geographic information system (GIS) software (QGIS), free government data, on-line plant databases, and published research data to derive a geographic information database at the county scale showing distribution of three pigweeds: Palmer amaranth, redroot pigweed, and waterhemp. Database queries (i.e., a request for information from a database) were used to demonstrate applications of the GIS for precision agriculture applications at the county level, such as tallying the number of counties affected by the pigweeds, identifying counties reporting glyphosate-resistant pigweed, and identifying cultivated areas located in counties with glyphosate-resistant pigweeds. This research demonstrated that free and open-source geographic information software such as QGIS has strong potential as a decision support tool, with implications for precision weed management at the county scale.
4. Light reflectance properties of cotton and Palmer amaranth canopies. To better implement control strategies for Palmer amaranth invasions in cotton production systems, consultants and producers need tools that can help them differentiate it from cotton. ARS scientists at Stoneville, Mississippi, trained a computer algorithm to differentiate Palmer amaranth plants from cotton plants based on light reflectance measurements of their canopies. The study focused on differentiating Palmer amaranth from cotton near-isogenic lines with bronze, green, and yellow colored leaves. Overall classification accuracies ranged from 77.8% to 88.9%. The highest accuracies were achieved for Palmer amaranth versus cotton yellow classification. Findings support further application of machine learning algorithms and light reflectance properties of plant canopies as tools for Palmer amaranth and cotton discrimination with potential application of this technology in site-specific weed management programs.
Review Publications
Huang, Y., Reddy, K.N., Fletcher, R.S., Pennington, D. 2018. UAV low-altitude remote sensing for precision weed management. Weed Technology. 32:2-6.
Fletcher, R.S., Reddy, K.N. 2018. Geographic information system for pigweed distribution in the US Southeast. Weed Technology. 32:20-26.
Fletcher, R.S., Turley, R.B. 2017. Employing canopy hyperspectral narrowband data and random forest algorithm to differentiate palmer amaranth from colored cotton. American Journal of Plant Sciences. 8:3258-3271.
Huang, Y., Lee, M.A., Nandula, V.K., Reddy, K.N. 2018. Hyperspectral imaging for differentiating glyphosate-resistant and glyphosate-susceptible Italian Ryegrass. American Journal of Plant Sciences. 9:1467-1477.
Fisher, D.K., Huang, Y. 2017. Mobile open-source plant-canopy monitoring system. Modern Intrumentation. 6:1-13.
Fisher, D.K., Fletcher, R.S., Anapalli, S.S., Pringle III, H.C. 2017. Development of an open-source cloud-connected sensor-monitoring platform. Advances in Internet of Things. 8:1-11.
Lin, F., Zhang, D., Huang, Y., Wang, X., Chen, X. 2017. Detection of corn and weed species by the combination of spectral, shape and textural features. Sustainability. 9(8):1-14.
Zhang, J., Huang, Y., Liu, P., Yuan, L., Wu, K. 2017. Noise-resistant spectral features for retrieving foliar chemical parameters. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. 10(12):5368-5380.
