Progress 04/01/23 to 03/31/24
Outputs
Target Audience:Vegetable growers and allied industry; students and faculty, extension agents. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?One PhD student and one postdoc attended the ASABE (American Society of Agricultural and Biological Engineers) international ASABE conference in Omaha, Nebraska, USA, on July 8-12, 2023. They also improved their skills in image processing through a paid subscription to online courses. We also offered the Summer Undergraduate Research Experience (SURE) program (to provide undergraduate students with an extended experience through an 8-week program) to four students. The students work on several hands-on projects in collaboration with the PhD student and postdoc. PI Ampatzidis was able to attend several extension and outreach venues and presented the findings of this project (please see next section). How have the results been disseminated to communities of interest?Dr. Ampatzidis presented findings of this project in several extension, educational, and outreach venues including: Robotics, automation, and AI for specialty crops. Florida AgTech and AI Expo, Punta Gorda, FL, December 14, 2023. Steam weeding in sugarcane production - An alternative tool for weed management. Florida Crystals webinar, December 7, 2023. Potential technologies to maximize drone use. South Florida Veg Growers Meeting: Whitefly Management, Immokalee, FL, November 29, 2023. AI-enabled technologies for precision management of specialty crops. USDA ARS-Northeast Area retreat, November 14, 2023. Drones and AI for Precision Crop Management. Florida Ag Expo, Gulf Coast Research and Education Center, Wimauma, FL, November 9, 2023. Artificial intelligenceenhanced technology for precision citrus production. Cold Hardy Citrus Field Day, Quincy, FL, October 26, 2023. Commercializing your AI: Lessons and challenges from spinning off AI companies. UF AI Days, Gainesville, FL, October 18, 2023. Collaborative research: NRI: High throughput multi-robot weed management for specialty crops. Foundational Research in Robotics (FRR) -National Robotics Initiative (NRI) Principal Investigators' (PI) Meeting, Arlington, Virginia, May 1-3, 2023. Advancing Florida's Specialty Crop Production and Harvesting using Mechanization, Precision Agriculture, Drones, Robotics, and Artificial Intelligence. UF ABE Centennial Seminar Series, Gainesville, FL, April 20, 2023. Artificial Intelligence and precision agriculture in vegetable production. Citrus Show, Vegetable Session, Fort Pierce, April 13, 2023. UAVs, Sensors, & Other Applied Research Technology. Research Center Administrators Society (RCAS) Meeting, Gulf Coast REC, February 6, 2023. AI-enhanced smart machinery for precision scouting and spraying. National Alliance of Independent Crop Consultants (NAICC) 2023 Annual Meeting and AG PRO EXPO, Nashville, TN, January 23-27, 2023. AI-enhanced pest and disease management in vegetable production. South Florida Vegetable Growers Meeting: Research Updates on Weed and Disease Management. Immokalee, January 12, 2023. What do you plan to do during the next reporting period to accomplish the goals?In the next year, we will implement the updated machine vision system with the new camera and evaluate it for achieving high-speed detection at a high detection accuracy. We will also integrate the IMU with the GPS to improve the velocity readings for the robotic platform. We will evaluate the upgraded autonomous smart spraying system on vegetable beds and compare the spraying efficiency with conventional methods. The goal is to improve the spraying accuracy and achieve a low percentage of missed targets (aka., weeds) and off-target sprays using fine-tuning of the system.
Impacts
What was accomplished under these goals?
For the machine vision and AI-based weed detection system (Obj. 1), we tested and installed a new industrial camera with software control and manual control of image capture parameters. We analyzed the image dataset (several weed categories and three crops) and performed class balance by collecting more images and by using image augmentation techniques. The new dataset was added to the old dataset and overrepresented and non-relevant classes were removed. The AI models were retrained using the latest YOLO v8 models (nano, small, and medium) and their performance was evaluated in all the classes. For the smart spraying system, experiments were conducted to evaluate the system's pressure fluctuation with pressure and flow rate variations. Three more pumps were added to the system based on the tests to supply a higher flow rate. An updated control system was designed and implemented to maintain a stable application pressure in the spraying system under nozzle activation. We also designed and performed experiments, in a low-cost setup, to analyze the spray area, spray length, and width variation with application pressure, spray height, wind, and vehicle speed, in order to select the best nozzles and configuration for the smart sprayer. The optimum range of application pressure, spray height, and vehicle speed was determined for a narrow cone nozzle. We developed a newer robotic platform with a larger wheelbase and four independent motorized wheels (Obj. 2). The electrical and electronic component layout was redesigned, and a new larger capacity battery and multi-step redundancies were added to provide multiple layers of electrical safety. The robotic spraying system was evaluated in vegetable fields. For thevehicle routing problem (Obj. 3), we further improved and evaluated the developedcapacitated arc routing model to treat various types of weeds using a fleet of autonomous vehicles in large fields. The uncertainty of weed distributions in the field is considered in this optimization. The weed distribution in the field is given by UAV collected data (from the weed detection model developed in the previous year). This model provides better results compared to the optimization models developed in previous years.
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
Vijayakumar V., Ampatzidis Y., Schueller J.K., Burks T., 2023. Smart spraying technologies for precision weed management: a review. Smart Agricultural Technology, 6, 100337, https://doi.org/10.1016/j.atech.2023.100337.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
Ampatzidis Y., 2023. Emerging and advanced technologies in agriculture. Link (Linking Industry Networks through Certifications; High School Teachers Training) Conference, Daytona Beach, FL, October 10-12, 2023.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
Ampatzidis Y., 2023. AI and Extension. Possibilities and Challenges. 2023 SR-PLN Middle Managers Conference, Next Generation: Evolving the Extension Enterprise, Orlando, FL, August 22-24.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
Vijayakumar V., Ampatzidis Y., Silwal A., Kantor G., 2023. 2023. Development of a machine vision and spraying system of an autonomous robotic sprayer for specialty crops. ASABE Annual International Meeting, Omaha, Nebraska, USA, July 8-12, 2023.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
Ampatzidis Y., 2023. Solutions to critical issues facing field and specialty crop production. Integrative Precision Agriculture Local Solutions Through Global Advances International Conference, Athens, Georgia, May 18-19, 2023.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
Ampatzidis Y., 2023. Agrifood systems in a Circular Economy Framework: Unlocking the Future . 11th Agrotechnology Conference, American Hellenic Chamber of Commerce, Thessaloniki, Greece, February 17, 2023.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
Ampatzidis Y., 2023. AI-enhanced smart machinery for precision scouting and spraying. Annual Meeting and Ag Expo of the National Alliance of Independent Crop Consultants (NAICC), Nashville, TN, January 23-27, 2023.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
Vijayakumar, V., Ampatzidis, Y., Silwal, A., Kantor, G., 2023. Specialty crop-specific robotic precision smart sprayer based on machine vision and PWM-controlled spraying system. 2nd AI in Agriculture Conference, April 17-19, 2023, Orlando, FL.
|
Progress 04/01/22 to 03/31/23
Outputs
Target Audience:Vegetable growers and allied industry; students and faculty, extension agents. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?One PhD student and one postdoc attended the Florida ASABE (American Society of Agricultural and Biological Engineers) and the international ASABE conference in Houston, TX. We also offered the Summer Undergraduate Research Experience (SURE) program (to provide undergraduate students with an extended experience through an 8-week program) to four students. The students work on several hands-on projects in collaboration with the PhD student and postdoc. PI Ampatzidis was able to attend several extension and outreach venues and presented the findings of this project (please see next section). How have the results been disseminated to communities of interest?Dr. Ampatzidis presented findings of this project in several extension, educational, and outreach venues including: AI in Health, Engineering, and the Sciences. Panel discussion at the UF AI Days event. Gainesville, FL, October 28, 2022. AI-enhanced smart farming. The 6th IEEE UV2022, International Conference on Universal Village, Virtual and main venue in Boston, USA, October 22-25, 2022 (Keynote Speaker). Robotics and artificial intelligence in Agriculture. ABE2012C: Intro to Biological Engineering (Guest Lecture), UF. October 12, 2022 AI-enhanced precision management in specialty crops. USDA National Agricultural Statistics Service, Research and Development Division (virtual seminar). September 29, 2022. Robotics and AI in specialty crop production. 4-H robotics team, Exploding Bacon (virtual seminar). September 26, 2022. Artificial intelligence in Agriculture. BSC2930 Frontiers in AI course (virtual lecture), UF, September 20, 2022. Introduction to Agricultural and Biological Engineering. Freshman Leadership Engineering Group (virtual event), UF, September 19, 2022. Remote sensing, AI, and automation in digital agriculture. International Smart Ag Meeting (virtual), China, July 15, 2022 How artificial intelligence can enhance precision agriculture and address climate change issues. UF/IFAS TREC Seminar Series, Homestead, FL, April 14, 2022. How artificial Intelligence can enhance agriculture production and address climate change issues. Volo Climate Correction Conference, Orlando, FL, March 17, 2022. Putting research and technological advances in the hands of stakeholders what we should do to convert research to practice (panel discussion). Envisioning 2050 in the Southeast: AI-driven innovations in agriculture. Auburn, AL, March 9-11. Emerging technologies and AI in precision agriculture. UF/IFAS In-Service training 32072, ANR-AIEET IST Module 1 (virtual lecture), February 28, 2022. Artificial intelligence for precision agriculture. UF/IFAS In-Service training 32032, New Technology for Commercial Crop Production (virtual lecture), February 23, 2022. AI applications in horticulture. Guest Lecture, Hort 3213 Fruit and Nut Production, Oklahoma State University, February 16, 2022. Artificial intelligence and smart farming. UF ABE Biocomplexity Engineering group seminars (webinar), January 18, 2022. What do you plan to do during the next reporting period to accomplish the goals?In the next year, we plan to train the weed classification models with different networks to improve detection accuracy. The goal is to identify the best possible AI model(s) after comparing more algorithms and networks. We will also re-design the PWM-controlled spraying system to improve efficiency. We will test new nozzles and conduct spray characteristic tests in a controlled environment and in the field. We have already developed an experimental design to evaluate the integrated system on vegetable fields. Finally, we will Integrate the smart spraying system with the autonomous robot developed by our Co-PI (Carnegie Mellon University) and evaluate it in vegetable fields.
Impacts
What was accomplished under these goals?
For the machine vision and AI-based weed detection system (Obj. 1), we trained the classification model with new data and compared two state-of-the-art object detection algorithms (YOLOv7 and Faster RCNN) and determined the performance of each for each class under consideration. We have also built a new remote-controlled prototype robot for initial testing (Obj. 2). This prototype is equipped with the spraying system, control box, pumps, camera, and GPS to perform field trials. For thevehicle routing problem (Obj. 3), we evaluated the developedcapacitated arc routing model to treat various types of weeds using a fleet of autonomous vehicles in large fields. The uncertainty of weed distributions in the field is considered in this optimization. The weed distribution in the field is given by UAV collected data (from the weed detection model developed in the previous year). This model provides better results compared to the optimization models developed in previous years.
Publications
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Vijayakumar V., Ampatzidis Y., Silwal A., Kantor G., Burks T., 2022. Machine vision and AI-equipped multi-crop robotic smart sprayer for precision weed management of three weed types in vegetable crops. FASABE Annual Conference and Trade Show, Clearwater Beach, FL, May 19-22.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Vijayakumar V., Partel V., Ampatzidis Y., Silwal A., Kantor G., 2022. Precision weed management based on machine vision and AI integrated to an autonomous robotic smart sprayer for specialty crops. ASABE Annual International Meeting, Houston, Texas, July 17-20.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Vijayakumar V., Costa L., Ampatzidis Y., 2022. AI applications in Smart Agriculture (poster). UF/IFAS AI Summit, Gainesville, FL, June 21, 2022.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Vijayakumar V., Ampatzidis Y., Silwal A., Kantor G., 2022. Machine vision and AI precision weed management using an integrated autonomous robotic smart sprayer for specialty crops. Envisioning 2050 in the Southeast: AI-driven innovations in agriculture. Auburn, AL, March 9-11 (poster).
- Type:
Journal Articles
Status:
Under Review
Year Published:
2023
Citation:
Vijayakumar V., Ampatzidis Y., Schueller J.K., Burks T., 2023. Smart spraying technologies for precision weed management: a review. Computers and Electronics in Agriculture.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2023
Citation:
Vijayakumar V., Ampatzidis Y., Silwal A., Kantor G., 2023. Specialty crop-specific robotic precision smart sprayer based on machine vision and PWM-controlled spraying system. AI in Agriculture Conference: Innovation and Discovery to Equitably meet Producer Needs and Perceptions, Orlando, FL, April 17-19, 2023.
|
Progress 04/01/21 to 03/31/22
Outputs
Target Audience:Vegetable growers and allied industry; students and faculty, extension agents. Changes/Problems:Because of Covid-19 and supply chain issues, there were several delays on material and equipment delivery. What opportunities for training and professional development has the project provided?Two PhD students and one postdoc attended seminars offered by NVIDIA on the use of embedded programming for real time AI applications. They also submitted three papers and attended the ASABE (American Society of Agricultural and Biological Engineers) virtual conference in July 12-16, 2021. We also offered the Summer Undergraduate Research Experience (SURE) program (to provide undergraduate students with an extended experience through an 8-week program) to four students. The students work on several hands-on projects in collaboration with the PhD students and postdoc. PI Ampatzidis was able to attend several extension and outreach venues and presented findings of this project (please see next section). How have the results been disseminated to communities of interest?Dr. Ampatzidis presented findings of this project in several extension, educational, and outreach venues including: Emerging technologies for specialty crops. LaBelle Rotary Club, LaBelle Fl, November 16, 2021. Smart and data-driven farming. ABE2012C: Intro to Biological Engineering (virtual lecture), October 6, 2021. How artificial intelligence can enhance precision agriculture. Institute for Learning in Retirement, Life-Long Learning programs at Oak Hammock (virtual lecture), October 5, 2021. Artificial intelligence in agriculture. UF ABE Florida Farm Bureau Board, UF ABE, Gainesville, September 22, 2021. Smart and data-driven agriculture. UF Frontiers of AI (virtual lecture), September 21, 2021. Artificial intelligence in crop production. UF Foundation National Board Assembly (virtual event), September 17, 2021. Emerging technologies for precision management in vegetables. SWFREC Vegetable seminar, Immokalee, May 28, 2021. AI applications for Florida crop production. AFIS AI Forum, Virtual Seminar, May 26, 2021. Emerging Technologies and AI for BMP. UF/IFAS BMP Virtual Summit. May 18, 2021. Artificial Intelligence in Agriculture. UF AI Town Hall, May 13, 2021. Artificial intelligence for precision agriculture applications. UF/IFAS Extension Symposium, May 3-7, 2021. Smart sprayer utilizing sensor fusion and AI. Ag Tech Expo, Immokalee, April 9, 2021. Artificial Intelligence in Agriculture. Florida Senators, UF President, and Board of Trustees. February 25, 2021. Emerging Technologies and AI for BMP. UF/IFAS Water Wednesdays. February 24, 2021. Artificial Intelligence in Agriculture. UF EGN1935-FOAI (31730) - Freshman Engineering: Frontiers of AI (virtual lecture). February 9, 2021. Applications of Artificial Intelligence in Precision Agriculture. Central District Ag BMP virtual meeting. February 3, 2021. Drones, artificial intelligence, and the future of pest management in vegetable crops. Annual vegetable growers virtual meeting. The Oregon processed vegetable commission. January 25, 2021. AI in agriculture. UF/IFAS Extension Associate Deans AI Retreat, January 12, 2021. What do you plan to do during the next reporting period to accomplish the goals?In the next year, we plan to collect more images of weeds and plants during the initial growing phases to balance the dataset and build a customized weed detection model. We plan to improve the weed detection model's precision and recall to >90%. For the smart sprayer, activities will include the improvement of the spraying resolution/accuracy. Several nozzles are tested for this purpose. We plan to evaluate the integrated system during Fall 2022. For Objective 3, we plan to continue working on Tasks 1-4. For Task 3, we plan to incorporate the more recent weed distribution information estimated from the operating vehicles into the model. We will explore appropriate modeling approach such as online optimization and design efficient solution methodologies.
Impacts
What was accomplished under these goals?
For the machine vision and AI-based weed detection system (Obj. 1), we have collected 3,018 images of three different weed types and target plants (dataset of 5,662 high resolution images total). An initial AI-based weed detection model was created on these images. To develop a unmanned aerial vehicle (UAV) and AI-based weed detection, classification, and distribution system, we collected data at 35, 60, and 100 ft from three different locations. Weed distribution estimation using image processing techniques applied to these images. This information will be used for the fleet optimization model (Obj. 3) Furthermore, we improved the sprayer's functionality and debugged the CAN BUS communication issues. For thevehicle routing problem (Obj. 3), we previously developeda capacitated arc routing model to treat various types of weeds using a fleet of autonomous vehicles in large fields. In this year, we further consider the uncertainty of weed distributions in this problem.We first provide a linear vehicle routing model to formulate the problem with deterministic weed distributions. Based on this model, we propose a robust optimization model to incorporate the uncertainty of weed distributions by using a set, call uncertainty set, that includes all possible realizations of weed distributions. The objective of the robust model becomes seeking a route feasible to all weed distribution realizations in the uncertainty set with the minimum cost. We utilizethe estimation of weed distributions fromUAVsin the construction of uncertainty sets. Numerical results demonstrate the need of using the robust approach and show that the proposed model leads to a better out-of-sample performance.
Publications
- Type:
Book Chapters
Status:
Published
Year Published:
2022
Citation:
Kakarla S.C., Ampatzidis Y., Park S., Adosoglou G., and Pardalos P., 2022. Emerging sensing technologies for precision agriculture. Book: Information and Communication Technologies for AgricultureTheme I: Sensors. Springer Nature, ISBN: 978-3-030-84143-0.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2022
Citation:
Zhu Z., Ampatzidis Y., Silwal A., Pardalos P., 2022. Robust route planning for autonomous vehicles. 2022 ASABE Annual International Meeting, July 17-20, 2022
- Type:
Other
Status:
Submitted
Year Published:
2022
Citation:
Vijayakumar V., Ampatzidis Y., Silwal A., Kantor G., 2022. Machine vision and AI precision weed management using an integrated autonomous robotic smart sprayer for specialty crops. Envisioning 2050 in the Southeast: AI-driven innovations in agriculture. Auburn, AL, March 9-11 (poster).
- Type:
Conference Papers and Presentations
Status:
Submitted
Year Published:
2021
Citation:
Adosoglou G., Park S., Ampatzidis Y., Pardalos P., 2021. A high-level task planning of autonomous robots with multi-dimensional loading constraints for precision weed management under field variability. 2021 Virtual ASABE Annual International Meeting, July 11-14, 2021, ASABE Paper No. 2100426, doi: 10.13031/aim.202100426.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2021
Citation:
Vijayakumar V., Partel V., Ampatzidis Y., Silwal A., Kantor G., 2021. Autonomous smart sprayer for precision weed management using machine vision and AI. 2021 Virtual ASABE Annual International Meeting, July 11-14, 2021.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2021
Citation:
Partel V., Costa L., Ampatzidis Y., 2021. Smart citrus tree sprayer using sensor fusion and artificial intelligence. 22021 Virtual ASABE Annual International Meeting, July 11-14, 2021, Paper Number: 2100525, doi:10.13031/aim.202100525.
|
Progress 04/01/20 to 03/31/21
Outputs
Target Audience:Vegetable growers and allied industry; students and faculty, extension agents. Changes/Problems:Because of Covid-19, the University of Florida's Southwest Florida Research and Education Center was closed for several months, and we were not able to offer the Summer Undergraduate Research Experience (SURE) program (to provide undergraduate students with an extended experience through an 8-week program). We plan to offer this project in Years 2-4. What opportunities for training and professional development has the project provided?In the first year of this project, two PhD students were hired at University of Florida to work on this project. The students were able to attend seminars offered by NVIDIA on the use of embedded programming for real time AI applications. They have also submitted two papers and plan to attend the ASABE (American Society of Agricultural and Biological Engineers) virtual conference in July 12-16. PI Ampatzidis was able to attend several extension and outreach venues and presented findings of this project (please see next section). How have the results been disseminated to communities of interest?Dr. Ampatzidis presented findings of this project in several extension, educational, and outreach venues including: Artificial Intelligence in Agriculture. Florida Senators, UF President, and Board of Trustees. February 25, 2021. Emerging Technologies and AI for BMP. UF/IFAS Water Wednesdays. February 24, 2021. Artificial Intelligence in Agriculture. UF EGN1935-FOAI (31730) - Freshman Engineering: Frontiers of AI (virtual lecture). February 9, 2021. Applications of Artificial Intelligence in Precision Agriculture. Central District Ag BMP virtual meeting. February 3, 2021. Drones, artificial intelligence, and the future of pest management in vegetable crops. Annual vegetable growers virtual meeting. The Oregon processed vegetable commission. January 25, 2021. AI in agriculture. UF/IFAS Extension Associate Deans AI Retreat, January 12, 2021. Smart technologies and AI for best management practices. IFAS Certified Crop Advisory (CCA) CEU training program. Theme: Crop Management and Soil & Water. Recorded presentation via the UF CANVAS, November 2020. Moderator and in the Roundtable at the Florida-Israel Agriculture Innovation Summit (Ag Summit for robotics, drones, and AI research), Webinar, November 18, 2020. UAVs and AI in Agriculture. Lecture at Precision Agriculture Class, University of Missouri, Webinar, November 3,2020. AI and emerging technology in horticulture. IST at the FSHS Conference, Webinar, October 20, 2020. AI applications in agriculture. UFInformaticsInstitute, Webinar, October 14, 2020. AI in Agriculture. ABE2012c-Introduction to Biological Engineering. Guest Lecture, September 30, 2020. AI applications in agriculture. UPL North America R&D Innovation Webinar. September 29, 2020. UAV-based plant nutrient analysis and a novel smart spraying technology. Virtual Citrus Meeting, September 15, 2020. AI in Agriculture. UF ABE Webinar "HWCOE Artificial Intelligence Initiative Vision", Round Table, September 11, 2020. Precision and smart spraying technology for vegetables. Vegetable Growers Meeting, SWFREC, Immokalee, August 20, 2020. Emerging technologies for vegetables. Vegetable and Specialty Crop Virtual Meeting, August 13, 2020. Precision and variable rate spraying and fertilizer applications. Virtual Ag BMP summit (Webinar), June 16, 2020. Emerging technologies for precision management in vegetables. SWFREC Vegetable Webinar, May 28, 2020. Partel V., and Ampatzidis Y., 2020. Machine vision and deep learning techniques for precision weed management. Precision Agriculture Webinar Series, Alabama Extension, May 21, 2020. What do you plan to do during the next reporting period to accomplish the goals?In Year-2, we plan to develop and in-field evaluate the first prototype of the autonomous and smart spraying vehicle. The first prototype of the smart spraying system (attached on an ATV) has already been developed. The machine vision and AI-based system will be developed further to increase weed detection and spraying accuracies. We also plan to develop further the intelligence multi-robot coordination system and optimize spraying applications. Finally, we will start collecting data to develop a feasibility study for the proposed autonomous spraying system.
Impacts
What was accomplished under these goals?
Most available conventional sprayers apply agrochemicals uniformly, even though distribution of weeds is typically patchy, resulting in wastage of valuable compounds, increased costs, crop damage risk, pest resistance to chemicals, environmental pollution, and contamination of produce. Minimizing the negative impacts of agrochemicals (and spraying technologies) is a major global societal challenge. This project tackles the above problem by developing an autonomous multi-crop robotic platform equipped with a machine vision and artificial intelligence (AI) based precision sprayer to distinguish target weeds from non-target objects (e.g., vegetable crops) and precisely spray on the desired target. The precision sprayer is targeted to simultaneously treat 3 different types of weeds -vine weeds, broadleaf, and grass weeds- for specialty crops such as tomato and pepper. The first prototype was developed and utilizes 22 spray nozzles controlled by 12V solenoid valves, with a response time of less than 50 ms, covering a total work length of 1.3 m. A combination of a LUCID Triton 5.0 MP RGB camera with an intel RealSense depth camera (D435) was used for capturing RGB images and depth information. The images were processed by an NVIDIA Xavier GPU running a YOLO V4 object detector, which is a state-of-the-art open source neural network. This smart spraying system achieved an overall precision of 71% and a recall of 78% (for plant detection and target spraying accuracy) when tested with real plants, and 91% accuracy and recall with artificial plants. The next prototype is expected to achieve a significantly higher overall precision and recall and would be integrated to the autonomous robotic platform, making it a novel technology in precision weed management. The autonomous robotic platform is under development and it will use RTK-GPS and LIDAR for navigation in the field. We are also currently investigating a combination of end-to-end deep learning based steering regressor with Generative Adversarial Network (GAN) for autonomous navigation. Furthermore, a task planning algorithm was developed to optimize spraying application for the fleet of weeding robots. This optimization algorithm considers multiple vehicles carrying multiple types of herbicides to treat various types of weeds in a field by taking into account the need to replenish, as well as the time cost associated with it and the variability of the field. We formulated this problem as an extension of the capacitated arc routing problem (CARP), where we considered multiple trips to and from the refilling unit, the time required for replenishment, as well as the multiple commodities being transferred. We developed two mathematical models, one in which optimizes only for the total distance traveled among all vehicles, while the second model also optimizes the maximum time spent for completing the task. We demonstrated solutions to vehicle routing problems for a weed treatment operation using three types of agrochemicals and a fleet of autonomous precision sprayers.
Publications
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2021
Citation:
Vijayakumar V., Partel V., Ampatzidis Y., Silwal A., Kantor G., 2021. Autonomous smart sprayer for precision weed management using machine vision and AI. 2021 Virtual ASABE Annual International Meeting, July 11-14, 2021.
- Type:
Book Chapters
Status:
Accepted
Year Published:
2021
Citation:
Kakarla S., Ampatzidis Y., Park S., Adosoglou G., Pardalos P., 2021. Emerging technologies for precision agriculture. Information and Communication Technologies for Agriculture Theme I: Sensors, EFITA. Springer Nature, Switzerland.
- Type:
Other
Status:
Other
Year Published:
2020
Citation:
Partel V., and Ampatzidis Y., 2020. Machine vision and deep learning techniques for precision weed management. Precision Agriculture Webinar Series, Alabama Extension, May 21, 2020 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2020
Citation:
Ampatzidis Y., 2020. Emerging technologies for precision management in vegetables. SWFREC Vegetable Webinar, May 28, 2020 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2020
Citation:
Ampatzidis Y., 2020. Precision and variable rate spraying and fertilizer applications. Virtual Ag BMP summit (Webinar), June 16, 2020 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2020
Citation:
Ampatzidis Y., 2020. Emerging technologies for vegetables. Vegetable and Specialty Crop Virtual Meeting, August 13, 2020 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2020
Citation:
Ampatzidis Y., 2020. Precision and smart spraying technology for vegetables. Vegetable Growers Meeting, SWFREC, Immokalee, August 20, 2020 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2020
Citation:
Ampatzidis Y., 2020. AI applications in agriculture. UPL North America R&D Innovation Webinar. September 29, 2020 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2020
Citation:
Ampatzidis Y., 2020. AI applications in agriculture. UF Informatics Institute, Webinar, October 14, 2020 (invited talk).
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2021
Citation:
Adosoglou G., Park S., Ampatzidis Y., Pardalos P., 2021. A high-level task planning of autonomous robots with multi-dimensional loading constraints for precision weed management under field variability. 2021 Virtual ASABE Annual International Meeting, July 11-14, 2021.
- Type:
Other
Status:
Other
Year Published:
2020
Citation:
Ampatzidis Y., 2020. AI and emerging technology in horticulture. IST (In Service training) at the FSHS Conference, Webinar, October 20, 2020 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2020
Citation:
Ampatzidis Y., 2020. Smart technologies and AI for best management practices. IFAS Certified Crop Advisory (CCA) CEU training program. Theme: Crop Management and Soil & Water. Recorded presentation via the UF CANVAS, November 2020 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2021
Citation:
Ampatzidis Y., 2021. Drones, artificial intelligence, and the future of pest management in vegetable crops. Annual vegetable growers virtual meeting. The Oregon processed vegetable commission. January 25, 2021 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2021
Citation:
Ampatzidis Y., 2021. Applications of Artificial Intelligence in Precision Agriculture. Central District Ag BMP virtual meeting. February 3, 2021 (invited talk).
- Type:
Other
Status:
Other
Year Published:
2021
Citation:
Ampatzidis Y., 2021. Emerging Technologies and AI for BMP. UF/IFAS Water Wednesdays. February 24, 2021 (invited talk).
|
|