Progress 03/15/23 to 03/14/25
Outputs
Target Audience: Researchers: students, scholars, faculty, extension educators, crop protection network Industry (crop protection, robotics) Phenomics Advisory Board: Bayer Crop Sciences, Corteva, Solinftec Farmers: Indiana, Pennsylvania, Honduras (including exporters to U.S.) Staff and faculty at Purdue and Penn State Changes/Problems: SCDA v1 did not work as anticipated, and we had to develop a new algorithm. Although full modeling based on data extracted from imagery is still pending, the models based on data collected visually (Lane et al., 2025) will soon be used to characterize tar spot epidemics based on digital data extracted from imagery. We faced a notable challenge with the SCDA algorithm's internal deployment on the Plant Village Nuru App. To address this issue, the Cruz Lab developed both semi-automated and automated methods (in collaboration with and industrial parner) for labeling and archiving images, which were then analyzed using the SCDA algorithm. What opportunities for training and professional development has the project provided? Research Skill Development Helped Da-Young Lee (Postdoc and later collaborator), Ojaswi Acharya (MSc student), Brenden Lane (PhD student), and Alex Acosta and Emanuel Santos (Visiting Scholars) conduct experiments data analysis and interpretation. Helped Alex Acosta develop an improved tar spot inoculation protocol from where images were acquired. Organized regular meetings in collaboration with Dr. Mohammad Jahanshahi (Co-PI) to discuss research methodologies and data interpretation associated with plant disease modeling. These meetings connected Dr. Jahanshahi's student (College of Engineering) with members of the Cruz lab (College of Agriculture). Guided MSc Student Ojaswi Acharya through the complexities of establishing adequate methods and practices prior to modeling. Tyler McFeaters, Education Program Specialist (Penn State) was mentored by Paul Esker and increased knowledge of working with field collected IoT-based images and using drones in sustainable plant disease management. Presentation and Communication Skills Encouraged graduate students Brenden Lane and Visiting Scholar Alex Acosta to present their research findings at national or international conferences. Conducted internal presentation sessions for students to practice public speaking and receive feedback. PhD student Brenden Lane presented his research findings at the annual Corn Disease Working Group Meeting. This meeting included attendees from extension and research programs representing both academia and industry, providing an opportunity to share critical insights and foster collaboration within the field. Technician Andrés Cruz-Sancan presented his proposal and findings in front of researchers at the International Epidemiology Workshop in Brazil. Technical Skills Enhancement Provided training on using advanced software tools (e.g., statistical analysis software, data visualization tools). Worked with Alex Acosta, Emanuel Santos, Andrés Cruz-Sancan, Emilia Castillo, and Ojaswi Acharya (student with Dr. Jahanshahi). Offered guidance on understanding and using concepts and techniques in plant disease epidemiology and crop protection. Supported Research Operations Administrator Andrés Cruz improving data quality management skills. Career and Academic Development Supported students in writing research paper drafts and grant proposals. Facilitated mentorship sessions on academic or industry career planning and networking strategies. Interdisciplinary Collaboration Involved students in collaborative projects with other research teams or industry partners. Arranged joint seminar with experts from other disciplines to broaden perspectives. Outreach and Extension Activities Engaged students (Purdue) and Education Program Specialist (Penn State) in community outreach programs related to agricultural education, digital and advanced agricultural systems, and disease management. Leadership and Project Management Appointed students to lead specific project components, fostering leadership and time management skills. Trained the in organizing and coordinating team activities. How have the results been disseminated to communities of interest?Through manuscript publications, oral presentations at the national and international level, direct engagement with stakeholders (including but not limited to the robotics and crop protection industry as well as farmers). Information was presented to just under 1,200 stakeholders during Penn State field and crops days and conferences. Tar spot was part of 26 talks over the past two years of the project. What do you plan to do during the next reporting period to accomplish the goals?
Nothing Reported
Impacts
What was accomplished under these goals?
What did you discover or learn? OBJECTIVE 1:After continuous engagement with multiple stakeholders including industry partners, farmers, product exporters, and researchers, and post completion of the NSF's Innovation Corps (I-Corps™) program, we concluded that timely disease management considering temporal disease thresholds based on data extracted from imagery arekey for decision making. Farmers and county Extension educators actively assisted in coproducing proximal imagery or sampling. For instance, during a general meeting with farmers attending a Digital Ag Showcase, 94% of respondents indicated that they believe that the successful implementation of this research has the potential to positively impact plant disease management and agriculture. The quality of the research in addressing plant disease phenotyping challenges was also approved by 94% of respondents. Proximal imagery and other datasets were generated in collaboration with Co-PI Esker and international collaborator Dr. Falconí. In addition, in 2024, we engaged with robotics company Solinftec and they provided additional datasets. So, it will be important to continue with these goals until models are fully implemented. OBJECTIVE 2:Experiments tested the ability to accurately simulate tar spot epidemics at the field level using data extracted from proximal imagery collected in multiple field environments (Lee et al., 2021, SCDA v1). However, the original SCDA v1 struggled with detecting and quantifying tar spot under field conditions. For that reason, we spent time developing SCDA v2 algorithm. Proximal RGB imagery, coupled with the enhanced Stromata Contour Detection Algorithm version 2 (SCDA v2), accurately and consistently quantifies corn tar spot severity. SCDA v2 integrates a convolutional neural network (CNN) to automate parameter optimization, significantly reducing detection errors compared to the previous version (SCDA v1). It exhibits high agreement with visual reference assessments across diverse field and greenhouse conditions, ensuring robust and transferable epidemiological metrics. Consequently, SCDA v2 streamlines the timely simulation and accurate measurement of tar spot epidemics at the field level, facilitating scalable and reliable data collection essential for effective disease management and modeling. Our results align well with Objective 2's emphasis on accurate and timely simulation of tar spot epidemics using proximal imagery. OBJECTIVE 3:Due to the limitations of SCDA v1, we initiated parallel development of a new algorithm (SCDA v2) along with targeted data collection. Datasets from fully automated IoT devices were gathered, and models were developed using visually collected data. We are currently testing the reliability of SCDA v2 and a newer algorithm developed in collaboration with a robotics partner in Indiana. In the coming months, these algorithms will be used to extract data from imagery collected in production-style fields. The extracted data will be compared to visual assessments to evaluate epidemic patterns. What activities or experiments did you conduct? Objective 1: engagement with Phenomics Advisory Board, growers, industry collaborators (robotics and crop protection companies), and peer researchers. On-farm research. Objective 2: Visual disease evaluations, imagery acquisition, data collection, sensor deployment, temporal acquisition of data, testing and development of algorithms, agreement analysis (accuracy, precision, reliability), partnership with robotics company, automation of data acquisition and storage, data curation. What results or key outcomes were achieved? Objective 1: Task 1: Identified and understood stakeholders' perceptions regarding tar spot management and proposed technology. Developed relationships and partnerships with diverse stakeholders (farmers, Extension educators, researchers). Identified critical gaps and practical adjustments needed in proposed solutions to meet real-world management needs. Consensus on epidemiological metrics relevant for timely disease management, including fungicide timing thresholds. Identified additional management scenarios based on canopy-level thresholds informed by data. Task 2: Selected and established diverse on-farm trial sites in Indiana and Pennsylvania. Collected comprehensive datasets, including weather data, crop growth indicators, crop health status, and yield information. Increased stakeholder involvement through participatory collection of proximal imagery and tissue samples, enriching datasets. Objective 2 Compared and validated proximal RGB imagery coupled with SCDA v1 and v2 for precise quantification of tar spot severity. Automated stromata detection and parameter optimization via integration of CNN into SCDA (v2), significantly enhancing accuracy and reducing error compared to SCDA v1. Developed epidemiological models (linear/non-linear) to effectively capture tar spot dynamics across multiple canopy levels and environments (based on visual assessments). Generated accurate epidemiological metrics (based on visual data), including disease intensity, disease onset, maximum severity, area under the disease progress curve (AUDPC), and apparent infection rates. Objective 3 Task 1: Deployed and field tested IoT devices for imaging to automatically and continuously monitor tar spot epidemics at scale. Enhanced the SCDA algorithm through CNN training using extensive historical datasets for improved accuracy and automation. Automatic upload of images and remote transmission to the cloud. Task 2: Automatic cloud uploading, secure storage (DataDepot), and accessibility of data collected by IoT devices. Extraction and modeling of comprehensive epidemiological parameters (visual intensity, onset, maximum severity, AUDPC, infection rates) across temporal scales. Agreement analyses to ensure reliability and accuracy of digital vs. visual assessments. Engagement with stakeholders through an NSF I-Corps™-inspired framework, ensuring technology innovation transitions effectively to real-world impact.
Publications
- Type:
Other Journal Articles
Status:
Published
Year Published:
2025
Citation:
Alexey Mikaberidze, C. D. Cruz*, Ayalsew Zerihun*, Abel Barreto, Pieter Beck, Roc�o Calder�n, Carlos Camino, Rebecca E. Campbell, Stephanie K.L. Delalieux, Fr�d�ric Fabre, Elin Falla, Stuart Fraser, Kaitlin M. Gold, Carlos Gongora-Canul, Fr�d�ric Hamelin, Dalphy O.C. Harteveld, Cheng-Fang Hong, Melen Leclerc, Da-Young Lee, Murillo Lobo Jr., Anne-Katrin Mahlein, Emily McLay, Paul Melloy, Stephen Parnell, Uwe Rascher, Jack Rich, Irene Salotti, Samuel Soubeyrand, Susan Sprague, Antony Surano, Sandhya D. Takooree, Thomas H. Taylor, Suzanne Touzeau, Pablo J. Zarco-Tejada, Nik J. Cunniffe. 2025. Opportunities and challenges in combining optical sensing and epidemiological modelling. Phytopathology. First Look https://apsjournals.apsnet.org/doi/epdf/10.1094/PHYTO-11-24-0359-FI
- Type:
Other Journal Articles
Status:
Published
Year Published:
2025
Citation:
Lane, B. 2025. Unraveling the dynamics of tar spot epidemics in corn fields: A DATA-DRIVEN FRAMEWORK FOR disease development and Surveillance. PhD dissertation https://hammer.purdue.edu/articles/thesis/_b_Unraveling_the_dynamics_of_tar_spot_epidemics_in_corn_fields_A_DATA-DRIVEN_FRAMEWORK_FOR_disease_development_and_Surveillance_b_/28908050
- Type:
Other Journal Articles
Status:
Published
Year Published:
2025
Citation:
Lee, D-Y., Na, D-Y., G�ngora-Canul, C., Jimenez-BeitiaV, F., Goodwin, S., Cruz-Sanc�n, A.T, Delp, E., Acosta, A.V, Lee, J-S., Falcon�, C., Cruz, C. D. 2025. Optimizing Tar Spot Measurement for Corn Health Analysis: A Deep Learning Approach Using Stromata Contour Detection Algorithm (SCDA) and Red-Green-Blue (RGB) Imaging. Plant Disease 109:73:83. https://doi.org/10.1094/PDIS-12-23-2702-RE
- Type:
Other Journal Articles
Status:
Published
Year Published:
2024
Citation:
G�ngora-Canul, C.P, Puerto, C.V, Jimenez, F.V, Telenko, D., Kleczewski, N., Rosas, J.C., Avellaneda, C., Sanders, A., Rodriguez, I., Goodwin, S., Henriquez-Dole, L., Fernandez-Campos, M., Lee, D., Cruz, Andr�s P.T, and Cruz, C. D.C 2024. Comparing tar spot epidemics in high-risk areas in the United States and Honduras. Canadian Journal of Plant Pathology. DOI: 10.1080/07060661.2023.2300077
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2025
Citation:
Lane, B. 2025. Unraveling Tar Spot Epidemic Dynamics in Indiana Corn Fields (2021-2024): Insights for Precision Management. 2025. Corn Disease Working Group Meeting. Pensacola, FL. Invited speaker.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2025
Citation:
Cruz, C. D. 2025. Role of Digital Technologies in the One Health Context. Purdue Institute for Digital and Advanced Agricultural Systems-Vet Med. West Lafayette, IN. Invited Panelist.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2025
Citation:
Cruz, C. D. 2025. Innovative Approaches in Phytopathometry, Epidemiology, and Microbiology: Focus on Corn Tar Spot Advancing Crop Biosecurity Through Science-Based Solutions. North Central Integrated Pest Management Center. Webinar. Invited speaker
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2024
Citation:
Lane, B. 2024. Characterizing Tar Spot Epidemics in U.S. Corn: A Medley of Models. I.E. Melhus Graduate Student Symposium. The American Phytopathological Society Annual Meeting. Memphis, TN.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2024
Citation:
Cruz, C. D. 2024. Using State-of-the-Art Imaging Methods for Epidemiological Purposes: The Role of Proximal and Remote Sensing Technologies. In: Innovative Approaches in Plant Disease Epidemiology: Navigating Challenges in a Technologically Driven Era. The American Phytopathological Society Annual Meeting. Memphis, TN. Invited Speaker and Session Organizer.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2024
Citation:
Cruz, C. D. 2024. Advancing Agriculture: Exploring strategic partnerships in plant pathology, digital tool development, and plant disease modeling. Bayer Crop Science. West, Lafayette, IN. Invited speaker.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2024
Citation:
Cruz, C. D. 2024. Promoting Agricultural Progress: Investigating Collaborative Alliances for Creating Digital Solutions and Forecasting Crop Diseases. Corteva Agriscience. Online. Invited speaker.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2024
Citation:
Cruz, C. D. 2024. Automated Plant Health Management. Solinftec Inc. Ara�atuba, Sao Paulo, Brazil. Invited speaker.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2024
Citation:
Cruz, C. D. 2024. New Technologies in Pest and Disease Monitoring in the Context of Climate Change. Regional Event: Climate Change and Its Impact on the Implementation of Sanitary and Phytosanitary Measures. San Pedro Sula, Honduras. Moderator and Speaker
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2024
Citation:
Cruz, C. D. 2024. Developing tools for high throughout plant disease phenotyping. Bayer Crop Science. Online. Invited speaker.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2024
Citation:
Cruz, C. D. 2024. Leveraging Advanced Imaging Techniques for Epidemiology: Exploring the Role of Proximal and Remote Sensing Technologies. Agronanal�tICA Workshop Organized by ICA. Bogot�, COLOMBIA. Keynote speaker.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2024
Citation:
Cruz, C. D. 2024. Modernizing Plant Disease Monitoring: Integrating UAV and Ground-Based Imaging in Epidemiological Research. International Epidemiology Workshop organized by the International Society of Plant Pathology and the Brazilian Society of Plant Pathology. Foz do Igua�u, BRAZIL. Invited speaker.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Centre County 2+2: Tar spot of corn. Six attendees. Bellefonte, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Centre County 2+2: The use of drones in agriculture. Six attendees. Bellefonte, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Mifflin County 2+2: Tar spot of corn. Eleven attendees. Lewistown, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Mifflin County 2+2: The use of drones in agriculture. Eleven attendees. Lewistown, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Franklin County Crops Day: Drones. Seventy-five attendees. Chambersburg, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Lebanon Crops Conference: Drones. Sixty attendees. Lebanon, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Lebanon Crops Conference: Tar spot. Sixty attendees. Lebanon, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Juniata County 2+2: The use of drones in agriculture. Eight attendees. Mifflintown, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Hortsdale Pesticide Meeting - Tar spot identification and management. Three hundred attendees. Chambersburg, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Clinton County 2+2: Tar spot of corn. Fifteen attendees. Mill Hall, PA
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Clinton County 2+2: The use of drones in agriculture. Fifteen attendees. Mill Hall, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Centre Crops Day: Tar spot of corn. Twenty-five attendees. Bellefonte, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Centre Crops Day: The use of drones in agriculture. Twenty-five attendees. Bellefonte, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Mercer Crops Day: The use of drones in agriculture. Twenty-five attendees. Mercer, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Virtual Crops Conference: The use of drones in agriculture. Twenty-seven attendees. Online
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Keystone Crops and Soils Conference: The 2024 disease year in review. 75 attendees. Grantville, PA.
- Type:
Other
Status:
Other
Year Published:
2024
Citation:
Esker, P. 2024. Soperstown Seed: Identifying foliar diseases of corn and soybean. 70 attendees. Bradford, PA.
- Type:
Other
Status:
Other
Year Published:
2025
Citation:
Esker, P. 2025. Centre County Crops Day: Tar spot of corn. 30 attendees. Bellefonte, PA.
- Type:
Other
Status:
Other
Year Published:
2025
Citation:
Esker, P. 2025. Washington County Crops Day: The 2024 disease year in review and drones in agriculture. 30 attendees. Washington, PA.
- Type:
Other
Status:
Other
Year Published:
2025
Citation:
Esker, P. 2025. Juniata County 2+2: The 2024 year in review and tar spot of corn. 5 attendees. Mifflintown, PA.
- Type:
Other
Status:
Other
Year Published:
2025
Citation:
Esker, P. 2025. Clinton County 2+2: The 2024 year in review and tar spot of corn. 10 attendees. Mill Hall, PA.
- Type:
Other
Status:
Other
Year Published:
2025
Citation:
Esker, P. 2025. Mifflin County 2+2: The 2024 year in review and tar spot of corn. 15 attendees. Lewistown, PA.
- Type:
Other
Status:
Other
Year Published:
2025
Citation:
Esker, P. 2025. Mid-Atlantic Fruit and Vegetable Conference: Precisi�n agricultura (Spanish program). 70 attendees. Hershey, PA
- Type:
Other
Status:
Other
Year Published:
2025
Citation:
Esker, P. 2025. Bedford County 2+2: Fungicide considerations in field crops. 150 attendees. Bedford, PA.
- Type:
Other
Status:
Other
Year Published:
2025
Citation:
Esker, P. 2025. Pennsylvania Agronomic Education Society: Fungicide considerations in field crops. 50 attendees. Boalsburg, PA.
|
Progress 03/15/23 to 03/14/24
Outputs
Target Audience: Stakeholders, and researchers with relevant experience Corn commodity board: Indiana Corn Marketing Council Phenomics Advisory Board: Composed of Companies: Corteva, Bayer, Beck's Hybrids, AgReliant Genetics Farmers: Indiana and Pennsylvania Staff and faculty at Purdue and Pennsylvania State University Changes/Problems:The project has made substantial progress by achieving key milestones, including the selection of farmer participants and the identification of research fields. Our on-farm research was carried-out efficiently, with the development of protocols for deploying IoT cameras across fields. These protocols specified camera placement and orientation, the schedule for data collection, and the completion of data pre-processing. However, we encounter a significant challenge integrating the SCDA v2 algorithm into the the Plant Village Nuru App ecosystem. The app was found to lack crucial features needed for organizing data and tracking disease intensity over time. To address this issue, the Cruz Lab devised both semi-automated and fully automated methods for labeling and archiving images, which were then analyzed using the SCDA algorithm. This ensured that the limitations of the app did not significantly hinder the progress of our project. Additionally, our preliminary observations suggest that images taken from close range may be more reliable than those obtained through standard IoT devices during the preprocessing stage. Moving forward, we will continue to process and analyzing the data already or to be collected and emphasize refining image acquisition protocols in the next reporting period. What opportunities for training and professional development has the project provided?Training Activities: Postdoctoral researcher Dr. Zhang and external collaborator Dr. Lee engaged in a structured discussions with technicians and Visiting Scholars from Purdue and Penn State to enhance the methodologies for image acquisition, aiming for greater consistency and precision across research projects. These training activities were comprehensive, encompassing preparatory meetings, discussion sessions during data collection, and debriefings afterwards to discuss outcomes and areas for improvement. Professional Development: Dr. Cruz has been recognized for his expertise and invited to engage in industry dialogues and present at academic conferences. These opportunities focused on sharing insights from recent research, exploring collaborative opportunities with industry partners, and discussing the latest trends and innovations in their field. How have the results been disseminated to communities of interest?One manuscript was submitted for publication and is accepted with revision. Results will be available for dissemination during year 2 of this project. What do you plan to do during the next reporting period to accomplish the goals?Improve image acquisition protocol. Complete final field and greenhouse work. Publish results, including improved SCDA (v2) algorithm. Complete epidemiological modeling efforts. Continue meetings with stakeholders (PAB in Indiana and others in Pennsylvania). Share final results with stakeholders.
Impacts
What was accomplished under these goals?
Objective 1: Task 1. Met with the Phenomics Advisory Board (PAB) and group was invited to participate in a Digital Showcase. We presented the hypothesis, our research vision, and preliminary results, identified gaps and established relationships with stakeholders. The goal of these first meetings was to understand perceptions of the value proposition design based on an assessment of how our proposed solutions match what they need. Task 2. We chose research sites in Indiana and Pennsylvania, through a combined effort by the Extension and research teams. We selected sites that provided a range of epidemics in a range of settings. Additionally, we collected imagery under greenhouse conditions from controlled inoculation experiments. We collected imagery, weather and crop growth data, and other indicators of crop health, and yield when available. We solicited Dr. Esker and his team and a collaborator overseas for assistance in coproducing proximal imagery to increase engagement and obtain sufficient data points. Objective 2: We included datasets from a variety of environmental settings, encompassing both proximal and IoT camera-based RGB imagery and detailed assessments of tar spot disease intensity on corn plants. This information, alongside data on weather conditions and crop growth stages, has been systematically organized and managed for comprehensive analysis. The algorithm SCDA version 2 (Stromata Contour Detection Algorithm) was used to quantify tar spot intensity by measuring both the number of stromata present and the percentage of the leaf area that is covered with stromata. Objective 3: Our team has performed data cleaning tasks to remove inaccuracies and extracted key insights from the collected imagery, focusing on indicators of crop health and disease presence. In the upcoming months, we will undertake additional analyses to further understand the potential of using data extracted from imagery to characterize disease progression.
Publications
- Type:
Journal Articles
Status:
Under Review
Year Published:
2025
Citation:
Lee, D-Y., Na, D-Y., G�ngora-Canul, C., Jimenez-Beitia, F., Goodwin, S., Cruz-Sanc�n, A., Depl, E., Acosta, A., Lee, J-S., Falcon�, C., Cruz, C. D. 20XX. Optimizing Tar Spot Measurement for Corn Health Analysis: A Deep Learning Approach Using Stromata Contour Detection Algorithm (SCDA) and Red-Green-Blue (RGB) Imaging. Plant Disease. Accepted with changes.
|