Digital Agriculture Center
Research team: Lazarovitch, N. (BGU), Levintal, E (BGU), Kool, D (BGU), Kira, O (BGU), Rachmilevitch, S (BGU), Furman, A (Technion), Fishbain, B (Technion), Linker, R (Technion), Kizel, F (Technion), Dubowski Y (Technion), Moshelion, M (HUJI), Friedlander, T (HUJI), Gavish, M (HUJI), Mau, Y (HUJI), Helman, D (HUJI).
Research Funding: The planning and budgeting committee. Israeli Center for Digital Agriculture
Students:
The Information Age has led to a transformative revolution in human history through tools such as machine learning (ML) and artificial intelligence (AI), which require abundant, high-quality, and reliable data. This requirement is particularly acute in agriculture, where complex and dynamic interactions exist within the soil-plant-atmosphere system. In response to the escalating global demand for food and environmental challenges, this proposal outlines a comprehensive initiative that will develop advanced functional phenotyping infrastructures and integrate cutting-edge sensors and AI technologies to foster agricultural research. The goal is to enhance plant productivity and sustainability by establishing state-of-the-art facilities for high-throughput, real-time monitoring and data collection of real-world farm environments, thereby bridging the gap between conventional practices and advanced technological solutions. The innovative aspect of this proposal lies in the cutting-edge facilities that will be established and used to generate high-throughput, high-quality, labeled data associated with key physiological plant responses under agricultural conditions, which is essential for training AI models. The project will focus on capturing multi-parametric plant diagnostic traits under frequently changing environmental conditions, offering real-time insights into the physiological and morphological responses of plants to such changes. This extensive monitoring of physiological characteristics, both within a controlled environment and under field conditions, will be integrated into learning systems based on actual plant responses, thereby incorporating reinforcement learning. Structured into the following five Work Packages (WPs), the project will target essential facets of plant–environment interactions and data-driven agricultural management: WP1: Design and construct an autonomous platform for real-time field crop monitoring using diverse sensors for comprehensive data collection in field conditions. WP2: Design and build a functional plant phenotyping system in a commercial-like greenhouse, featuring advanced feedback-irrigation, sensor integration, and imaging technologies to continuously monitor the instantaneous response of plants to varying growing conditions. WP3: Integrate, calibrate, and test sensors that will accurately monitor biological, physical, and chemical parameters within the soil–plant–atmosphere system. WP4: Build a centralized data server and develop ML and AI algorithms, focusing on generative AI and reinforced learning, which will allow predictive modeling and optimization of crop production based on multiple aspects of plant physiology and the dynamic interactions of plants with the soil and atmosphere. WP5: Disseminate the results to stakeholders, including universities, research centers, and the public, through workshops, open days, educational initiatives, and hackathons.
