Location: Northern Great Plains Research Laboratory
Project Number: 3064-21600-001-010-S
Project Type: Non-Assistance Cooperative Agreement
Start Date: Sep 1, 2022
End Date: Aug 31, 2027
Objective:
AM#2 additional objectives:
(6) Application of artificial intelligence (machine learning, ML) in the evaluation of field plant species composition and levels of grazing;
(7) Development of cattle head-mounted camera system (CHMCS) and pole-mounted camera system (PMCS) with machine vision (MV) technology to determine the plant species composition and preference by cattle.
AM#1
(1) Application of remote sensing imagery for forage and field crop biomass prediction; (2) cereal and broad leaf crops disease assessment through foliar image processing and other field image processing tools; and (3) assessment of circular bioeconomy (CBE) in crop and rangeland applicable to North Dakota.
(4) Application of single-board computer (SBE) for machine vision applications and machine learning in agriculture; (5) Development of circular bioeconomy (CBE) and net-zero agriculture (NZA) analysis frameworks.
Approach:
AM#2
Objective 6: (1) literature review on cattle grazing evaluation; (2) collect digital images of the plants and weeds in the field using hand-held digital (complete coverage of field) and UAV cameras; (3) collect ground truth samples; (4) development ML models and compare their efficiencies; (5) extend the digital library using data augmentation methods and compare their efficiencies; (6) “before” and “after” grazing images used for biomass removal evaluation; and (7) apply high-performance computing systems and develop related software routines and tools.
Objective 7: (1) development CHMCS and integrate with Raspberry Pi5; (2) parse the image data from CHMCS and identifying the species from the library; (3) apply the hand-held camera and UAV image data (to determine the plant species composition and preference; (4) evaluate grazing behavior from the CHMCS images; (5) develop full-scale and light-weight ML models and for computer and Raspberry Pi5 deployment; and (6) develop pole-mounted camera system (PMCS) to monitor animal behavior in the confined field.
AM#1
Objective 1: (1) collect historical remotely sensed imaging data applications; (2) perform a literature search of remote sensing in agricultural and rangeland scenarios; (3) develop assessment methods and determine their effectiveness; (4) develop and train ML models; and (5) develop web-based user-friendly tool.
Objective 2: (1) conduct a literature review of foliar diseases on cereal and broad leaf crops; (2) collect field images at different scales (leaf, subplot, and whole); (3) develop rapid disease assessment methods; (4) develop ML and deep learning (DL) models; (5) perform problem specific image processing tools such as weed identification, stand count, grazing evaluation, and LTAR relevant image processing; and (5) develop web-based tool for visualization and deployment.
Objective 3: (1) perform a literature search of CBE applicable to crop and forage production; (2) develop CBE systems analysis and apply to study cases with existing fields; (3) conduct scenario analysis to evaluate the benefits of CBE; and (4) develop tools to determine the benefits and visualize the results.
Objective 4: (1) finalize SLR review manuscript on the SBE applications in agriculture; (2) collect RP4 machine vision and ML model procedures and methodologies; (3) develop test programs implemented on RP4; (4) train machine vision algorithms on RP4; and (5) collect field images, train and test ML models on RP4 for agricultural applications.
Objective 5: (1) draft review of CBE in cropland agriculture; (2) perform SLR on the status of NZA in North American agriculture; (3) formulate framework to conduct “what-if” scenario analysis on CBE and NZA of ND crops; (4) develop tools to implement the initial framework.
