Progress 12/15/23 to 12/14/24
Outputs
Target Audience:• Faculty atAgAID member organizations • External research collaborators in academia and industry • International Partners - Organizations, agencies, and institutions from other countries • Students and educators in agriculture related programs at educational institutions in the USA and in Australia, Chile, and Netherlands • Specialty crop growers, both in the USA and in Australia, Chile, and Netherlands • Specialty crop agricultural farm managers and workers • Crop commissions and crop consultants • Water irrigation districts • State departments of agriculture and ecology • IT industry partners • AgTech and FarmTech industry partners • Federal funding agencies including USDA NIFA and NSF Changes/Problems:At the Merced campus, PhD student Gustavo Facincani Dourado graduated shortly after his visitation to Chile and was not able to fully finish the study. This work is now being undertaken by others and will be completed in Fall semester 2025. What opportunities for training and professional development has the project provided?The AgAID Institute contributed to the training and professional development of a variety of individuals. A summary of the numbers by the different categories is presented below. Please note that the reported numbers represent a lower bound on the number of trainees. • 1 undergraduate student 1 early-career project scientist • 3 PhD students (one of them was supported in Y2, the other two were supported in Y1) How have the results been disseminated to communities of interest?AgAID Institute's International Collaboration research has so far led to several peer-reviewed publications accepted or published or submitted for review in its first two years. These publications have appeared in leading AI and Ag conference and journal venues. The Institute members also organized various events such as workshops, and presentations at international conference venues aimed at the broader research community. A list of these outreach events from Year 2 is provided in the Other Products section. Various members of the Institute also presented AgAID International Collaboration related materials at several meetings, including regional, national and international conferences. What do you plan to do during the next reporting period to accomplish the goals?Merced-PUCC: Proceeding with the partnership with the Pontificia Universidad Catolica in Chile, in Fall 2025, UC Merced/AgAID staff and faculty will join the Global Change Center laboratory and complete development of scenarios of future crop suitability, identify at risk areas, and areas of opportunity for crop translocation. More specifically, the collaborative team will finish updating climate projections for winegrapes along the Pacific Coast of North and South America, to determine if finer resolution data significantly can alter our understanding of specialty crop viability in these regions. This work uses an ensemble machine learning approach to determine areas of suitability under future climate conditions. OSU-WUR: Using additional funding from the Washington Tree Fruit Research Commission, the OSU-WUR team will conduct joint robotic harvesting experiments at a commercial orchard in Prosser, WA in Fall 2025. WSU-UTS: The work for this task has been completed and our goals have been accomplished in Years 1 and 2 of the project.
Impacts
What was accomplished under these goals?
Accomplishments and progress are reported below by the different international collaborations. 1. Collaboration with AgAID-UC Merced team with PUCC-Chile: Viticulture is vital to the Americas' cultural heritage and generates significant agricultural income through high-quality wines. It also promotes sustainable practices and biodiversity preservation. Identifying suitable viticulture areas is essential for wine industry resilience in the face of climate change. Our approach employed species distribution modeling R package (sdm) to characterize the major winegrowing regions throughout the Americas, starting with American Viticultural Areas (AVAs) in California. We used vineyard data from LandIQ's 2021 land use mapping. We gathered global datasets with 20 bioclimatic, 46 edaphic, and 3 topographic variables from AVA vineyards extracted and/or calculated from the geodata R package. We tested these variables for multicollinearity and applied species distribution models using several statistical, machine learning and AI models such as GAM, GLM, GLMNET, GLMPoly, Maxent, RF, BRT, GBM, MARS, MaxLike, MLP, RBF, RPART, SVM, RangerRF, MDA, FDA, among others. Model accuracy was assessed using multiple replicates and performance metrics (e.g. AUC, Correlation, Deviance, and TSS). We also incorporated climate data for 2041-2060 to predict future agricultural suitability from 24 CMIP 6 Global Circulation Models. Preliminary results from the Napa Valley region reveal its unique suitability within California, which is going to significantly reduce later in the century. To refine predictions, we included edaphic and topographic data to complement climate-based models. Future work extends to Oregon, Washington, British Columbia, Baja California, and Chile, covering the Pacific Coast. We will also explore finer-resolution data (e.g., 30m- resolution DEM, edaphic characteristics for the conterminous US from NRCS inventory) and specific varietal suitability, possibly using growing degree days or a Winkler scale for region selection. In addition, we developed a partnership with USDA to work on two scientific papers on adaptation and mitigation strategies for vineyards. This information will aid growers and policymakers in making informed decisions about vineyard establishment, relocation, and the adoption of adaptive strategies to maintain wine production and quality in the coming decades. 2. Collaboration with AgAID-OSU team with WUR-Netherlands: During the first year of the project, one Robotics PhD student from Oregon State University (OSU) completed a three-month internship with collaborators at Wageningen University and Research (WUR). The focus of the research internship was modeling and shape reconstruction of dormant, woody plants, such as trees and shrubs, using 2D video. The outcomes of that work have been adopted by our WUR colleagues for robotic dormant pruning of currants. During the second year of the project (i.e. this performance period), our research focus transitioned to studies of in-hand sensing for robotic fruit harvesting. The OSU team fabricated and shipped to WUR a copy of our apple proxy for use in lab-based studies in the Netherlands. Additionally, OSU graduate students and an undergraduate AgAID intern developed new robotic picking controllers and a novel algorithm for using force sensing to detect the moment a fruit is picked from the tree. The controllers and algorithm were successfully integrated with WUR's robotic apple harvester and evaluated during joint field experiments at the research orchard in Randwijk, Netherlands in September 2024. Analysis of the datasets collected during these experiments is ongoing. Other highlights from this performance period include a visit from a WUR graduate student to the OSU Robotics lab in Corvallis, OR in April 2024, and a seminar from the OSU Co-PI at WUR in September 2024. 3. Collaboration with AgAID-WSU team with UTS-Australia: During this reporting period, Manoj Karkee (Co-PI) traveled to Australia in December 2023 to exchange collaboration ideas and to share the work being conducted at WSU with researchers at the University of Technology Sydney (UTS). He also gave a research seminar to an audience of ~30 people on "Automation and Robotics in Tree Fruit Orchards". Martin Churuvija (WSU PhD Student) traveled to Australia as a visiting scholar (for 3 months) to work at UTS' Robotics Institute, leveraging Washington State University's (WSU) partnership with UTS. Before Martin's visit to Australia, our team met regularly with the UTS team led by Professor Alen Alempijevic to develop collaborative research plans on utilizing WSU's expertise on sensing and machine vision systems to address agricultural issues in Australia while also utilizing the expertise at UTS including deformable object reconstruction to help with robotic pruning project. During the visit, Martin was supervised by Professor Alempijevic, a leading researcher in robotic perception, where he received training in object-oriented programming and libraries, threading and synchronization, and the Robot Operating System (ROS) middleware. In addition, Martin contributed to upgrading a robotic system that used manipulators and depth cameras to detect and reconstruct partially occluded apples on planar canopy trees. Additionally, with guidance from Dr. Alempijevic, he enhanced a pose-versatile imaging system that models planar-canopy fruit trees for automated orchard operations (e.g., robotic pruning) for his research at WSU.
Publications
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2024
Citation:
A. Velasquez Lopez, O. Gehrke, C. Grimm, and J.R. Davidson, Air pressure servoing for robotic grasping with multi-cup suction grippers , in IEEE/RSJ Intl Conf. on Intelligent Robots and Systems (IROS) Workshop on: AI Robotics for Future Farming, Abu Dhabi, United Arab Emirates, Oct. 2024, 4pp.
|
Progress 12/15/22 to 12/14/23
Outputs
Target Audience: Faculty at all AgAID member organizations External research collaborators in academia and industry International Partners - Organizations, agencies, and institutions from other countries Students and educators in agriculture related programs at educational institutions in the USA and in Australia, Chile, and Netherlands Specialty crop growers, both in the USA and in Australia, Chile, and Netherlands Specialty crop agricultural farm managers and workers Crop commissions and crop consultants Water irrigation districts State departments of agriculture and ecology IT industry partners AgTech and FarmTech industry partners Federal funding agencies including USDA NIFA and NSF Changes/Problems:At the Merced campus, one international student had her student visa denied and could not join the lab in the Fall of 2023. What opportunities for training and professional development has the project provided?The AgAID Institute contributed to the training and professional development of a variety of individuals. A summary of the numbers by the different categories is presented below. Please note that the reported numbers represent a lower bound on the number of trainees. 1 early-career project scientist 3 PhD students How have the results been disseminated to communities of interest?AgAID Institute's International Collaboration research has so far led to 1 peer-reviewed publications accepted or published in its first year, and 6 more papers which are currently undergoing review. These publications have appeared in leading AI and Ag conference and journal venues. A complete list of publications along with their full citation information is provided in the Products section. The Institute members also organized various events such as workshops, and presentations at international conference venues aimed at the broader research community. A complete list of these outreach events is provided in the Other Products section. Various members of the Institute also presented AgAID International Collaboration related materials at several meetings, including regional, national and international conferences. These presentations are listed below. Presentations: Merced-PUCC: Facincani Dourado, G. Using water-energy systems models to better inform decision-making. Technical Presentation. Turlock Irrigation District. Turlock, California. September, 2023. Facincani Dourado, G., Sarwar, A., Abatzoglou, J., Medellin-Azuara, J., Viers, J. H. Applying Machine Learning to Sustain Current and Future Winegrowing in California. Poster presentation. AgAID Annual Meeting. Wenatchee, Washington. September, 2023. Facincani Dourado, G., Rheinheimer, D. E., Abatzoglou, J. T., Viers, J. H. California's climate whiplash: Uncertainty in hydropower production, agricultural and environmental water supplies. Poster presentation. AgAID Annual Meeting. Wenatchee, Washington. September, 2023. Facincani Dourado, G., Rheinheimer, D. E., Abatzoglou, J. T., Viers, J. H. California's climate whiplash: fluctuations and uncertainty in hydropower production, and agricultural and environmental water supplies. Oral presentation. Yosemite Hydroclimate Meeting. Yosemite National Park, California. October, 2023. Facincani Dourado, G., Viers, J. H. Climate, Soils and Landscape: Applying Machine Learning and Principles of Vinecology to Sustain Current and Future Winegrowing along the Pacific Coast of the Americas. Poster presentation. American Geophysical Union Fall Meeting. San Francisco, California. December, 2023. OSU-WUR: J.R. Davidson, Invited Talk, IEEE/RSJ Int'l Conf. on Intelligent Robots and Systems (IROS) Workshop on: Agricultural Robotics for a Sustainable Future, "Towards tactile perception for manipulation in orchards", Detroit, MI, October 2023 J.R. Davidson, Invited Talk, FIRA USA, "An overview of robotic pruning research at the AgAID Institute", Salinas, CA, September 2023 WSU-UTS: M. Karkee, Invited Talk, FIRA USA, "Ag Robotics Research at Washington State University", Salinas, CA, September 2023 Sapkota, R., Ahmed, D., Khanal, S., Mo, U. B. C., Whiting, M. D., & Karkee, M. Robotic Pollination of Apples in Commercial Orchards. 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) October 1 -5, 2023, Detroit, USA What do you plan to do during the next reporting period to accomplish the goals?Merced-PUCC: Proceed with the partnership with the Pontificia Universidad Catolica in Chile, through a UC Merced exchange student joining the Global Change Center laboratory to co-develop scenarios of future crop suitability to identify at risk areas and areas of opportunity for crop translocation. More specifically, to work on refined and updated projections of Hannah et al. 2013 for winegrapes along the Pacific Coast of North and South America, to determine if finer resolution data significantly can alter our understanding of specialty crop viability in these regions. OSU-WUR: We will evaluate our tree reconstruction algorithm in a commercial orchard (Prosser, WA) in Winter 2024. With new funding from the Washington Tree Fruit Research Commission, OSU recently expanded its collaboration with WUR to include robotic apple harvesting. A team from OSU will visit the Randwijk experimental orchard (Randwijk, Netherlands) in September 2024 to conduct joint field experiments with our apple harvesting gripper prototypes. WSU-UTS: We will assess the applicability of a sensing system developed at WSU to perform 3D reconstruction of apple and cherry trees in WA and in Australia as well as in assessing the growth rate of animals such as pigs. Note that the WSU student international visit at UTS is scheduled for Y2. Therefore most of the project's research and collaboration will be carried out in Y2.
Impacts
What was accomplished under these goals?
Accomplishments and progress are reported below by the different international collaborations. 1. Collaboration with AgAID-UC Merced team with PUCC-Chile: Viticulture is vital to the Americas' cultural heritage and generates significant agricultural income through high-quality wines. It also promotes sustainable practices and biodiversity preservation. Identifying suitable viticulture areas is essential for wine industry resilience in the face of climate change. Our approach employs species distribution modeling R package (sdm) to characterize the major winegrowing regions throughout the Americas, starting with American Viticultural Areas (AVAs) in California. We used vineyard data from LandIQ's 2021 land use mapping. We gathered global datasets with 20 bioclimatic, 46 edaphic, and 3 topographic variables from AVA vineyards extracted and/or calculated from the geodata R package. We tested these variables for multicollinearity and applied species distribution models using several statistical, machine learning, or AI models such as GAM, GLM, GLMNET, GLMPoly, Maxent, RF, BRT, GBM, MARS, MaxLike, MLP, RBF, RPART, SVM, RangerRF, MDA, FDA, among others. Model accuracy was assessed using multiple replicates and performance metrics (e.g. AUC, Correlation, Deviance, and TSS). We also incorporated climate data for 2041-2060 to predict future agricultural suitability from 24 CMIP 6 Global Circulation Models. Preliminary results from the Napa Valley region reveal its unique suitability within California, which is going to significantly reduce later in the century. To refine predictions, we included edaphic and topographic data to complement climate-based models. Future work extends to Oregon, Washington, British Columbia, Baja California, and Chile, covering the Pacific Coast. We will also explore finer-resolution data (e.g., 30m-resolution DEM, edaphic characteristics for the conterminous US from UC Davis' inventory) and specific varietal suitability, possibly using growing degree days or a Winkler scale for region selection. In addition, we developed a partnership with USDA to work on two scientific papers on adaptation and mitigation strategies for vineyards. This information will aid growers and policymakers in making informed decisions about vineyard establishment, relocation, and the adoption of adaptive strategies to maintain wine production and quality in the coming decades. 2. Collaboration with AgAID-OSUteam with WUR-Netherlands: The major activity for our international collaboration with Wageningen University and Research (WUR) this performance period was a graduate student exchange. One Robotics PhD student from Oregon State University (OSU) visited the WUR Agri-Food Robotics group (Wageningen, Netherlands) from April-June 2023. The purpose of the exchange was joint collaboration on robotic dormant season pruning, which is a signature project within the AgAID Labor thrust (crop: perennial fruit trees) and also a focus of the Dutch Next Fruit 4.0 project (crop: red currants). The OSU graduate student made substantial contributions to plant modeling and reconstruction during his 3-month exchange. Creating accurate 3D models of tree topology is an important task for tree pruning. The 3D model is used to decide which branches to prune and then to execute the pruning cuts. Previous methods for creating 3D tree models have typically relied on point clouds, which are often computationally expensive to process and can suffer from data defects, especially with thin branches. We created an algorithm for actively scanning along a primary tree branch, detecting secondary branches to be pruned, and reconstructing their 3D geometry using just an RGB camera mounted on a robot arm. Using a laboratory setup, we experimentally validated that our method is able to produce primary branch models with 4-5 mm accuracy and secondary branch models with 15° orientation accuracy with respect to the ground truth model. Our framework is real-time and can run up to 10 cm/s with no loss in model accuracy or ability to detect secondary branches. The results of this work were published in two papers that included authors from both OSU and WUR. 3.Collaboration with AgAID-WSUteam with UTS-Australia: The WSU team is working with a faculty member at the University of Technology, Sydney in defining specific research tasks for this collaborative project. These research activities will be carried out next year. A WSU Ph.D. student (Martin Churuvija) will be exchanged to UTS and spend 3 to 4 months conducting collaborative research activities on sensing and automation technologies for fruit crop and animal farming. Most of the work with this collaboration will be done in Y2 of the grant, with the international student visit starting in Y2.
Publications
- Type:
Journal Articles
Status:
Under Review
Year Published:
2023
Citation:
Facincani Dourado, G., Rheinheimer, D. E., Abatzoglou, J. T., Viers, J. H. Stress testing California's hydroclimatic whiplash: Challenges, trade-offs and adaptations in water management and hydropower generation. Water Resources Research.
- Type:
Journal Articles
Status:
Other
Year Published:
2024
Citation:
Parker, L., et al. Science-Backed Guidance for Climate Resilient Vineyards. American Journal of Enology and Viticulture. In preparation.
- Type:
Conference Papers and Presentations
Status:
Under Review
Year Published:
2024
Citation:
A. You, J. Hemming, C. Grimm, and J.R. Davidson, A real-time, hardware agnostic framework for close-up branch reconstruction using RGB data, submitted to the 2024 IEEE Intl Conf. on Robotics and Automation (ICRA), under review.
- Type:
Journal Articles
Status:
Other
Year Published:
2024
Citation:
Viers, J. H., Facincani Dourado, G. et al. Combining Machine Learning and Principles of Vinecology to Sustain Current and Future Winegrowing along the Pacific Coast of the Americas. In preparation.
- Type:
Conference Papers and Presentations
Status:
Under Review
Year Published:
2023
Citation:
A. You, J. Hemming, C. Grimm, and J.R. Davidson, Branch orientation estimation using point tracking with a monocular camera, in IEEE Intl Conf. on Robotics and Automation (ICRA) Workshop on: TIG-IV: Agri-Food Robotics-From Farm to Fork, London, United Kingdom, June 2023, 2pp.
- Type:
Journal Articles
Status:
Under Review
Year Published:
2023
Citation:
Bhattarai, U., Zhang, Q., & Karkee, M. (2023). Design, Integration, and Field Evaluation of a Robotic Blossom Thinning System for Tree Fruit Crops. Computers and Electronics in Agriculture
|