A new computer vision system, developed by a research team at Penn State, has the potential to revolutionize the monitoring of specialty crops in controlled environment agriculture (CEA) systems. These systems, which allow for soilless growing in greenhouses, promise year-round production of high-quality crops. However, to remain competitive and sustainable, these advanced agricultural methods require the integration of precision agriculture techniques. To address this need, the Penn State team created an automated system that continuously tracks plant growth and provides frequent, real-time data on crop conditions, enabling more informed crop management decisions.
Traditionally, monitoring crops in CEA systems is a time-consuming process that requires skilled personnel. The conventional methods of crop monitoring do not allow for frequent data collection, which is crucial to understand the dynamic growth patterns of plants throughout their life cycle. By introducing automated crop-monitoring systems, researchers can collect data continuously and more efficiently, ensuring that crop management is based on real-time information.
The core innovation of this research is the integration of the Internet of Things (IoT), artificial intelligence (AI), and machine vision technologies. The IoT, a network of interconnected devices that exchange data, plays a critical role in this system by linking sensors and other technologies to provide continuous monitoring. The AI component of the system, along with a specialized recursive image segmentation model, processes sequential high-resolution images captured at regular intervals. This model tracks the growth of individual plants by analyzing changes in leaf coverage area throughout their development.
The research team tested the system by monitoring baby bok choy, a leafy vegetable often used in controlled environment farming. However, the team believes this system can be applied to a variety of crops, making it a versatile tool for precision agriculture. By accurately isolating and tracking plant growth, the system generates precise data about the plants’ progress and needs, ensuring that farmers can make informed decisions about irrigation, nutrient management, and environmental control.
The team behind this innovation has focused on developing robotic and automated solutions for agriculture for over a decade. Their previous work includes creating technologies for crop picking, pruning, thinning, pollination, and other agricultural processes. The machine vision system employed in this study builds on these earlier innovations, providing a more advanced tool for monitoring crop growth.
Chenchen Kang, the lead researcher responsible for the AI programming and system design, played a crucial role in the project. Kang installed the sensors, collected and processed the data, and developed the methodology and AI models needed to train the computer vision system. His expertise in machine learning and agricultural engineering was essential to the system’s success.
This research represents a collaborative effort between agricultural engineers and plant scientists, with contributions from several experts in the field. The project is part of a larger federal initiative aimed at advancing the sustainability of indoor urban agricultural systems. Francesco Di Gioia, an associate professor of vegetable crop science, emphasized the importance of integrating diverse expertise to create innovative solutions for controlled environment agriculture. This interdisciplinary approach, he believes, will be critical in making these systems more efficient and sustainable in the long term.
The integration of precision agriculture technologies such as IoT, AI, and computer vision will allow for the automatic monitoring of crops, including estimating growth rates, monitoring environmental factors such as radiation, temperature, and humidity, and managing nutrient solutions. These advancements will help minimize inefficiencies in crop production, making controlled environment agriculture more competitive and sustainable. Furthermore, these technologies could lead to improved crop quality, including tailoring the nutritional profile of specialty crops.
As these technologies continue to evolve, they hold the potential to revolutionize the future of controlled environment agriculture, increasing food security and contributing to more efficient and sustainable farming practices.
