Revolutionizing Farming: Computer Vision System for Specialty Crops

Researchers at Penn State have designed an IoT and AI-enabled automated crop monitoring system to boost controlled environment agriculture within greenhouses. The system makes it possible for real-time tracking and recording of plant development, improving crop management more efficiently.

Controlled environment agriculture using soilless growing systems supports year-round cropping of specialty plants. Traditional monitoring of crops should be conducted by skilled personnel and does not promote easy access to information. The system developed by the Penn State researcher optimizes efficiency with regards to ongoing plant growth data, improving decision making.

A definition of the technology is given by a Computers and Electronics in Agriculture report. The system integrates IoT, AI, and computer vision to monitor plant growth during the entire crop cycle. IoT connects physical devices with sensors and embedded software, providing data exchange and real-time monitoring.

The principal contribution of the work is a recursive image model of segmentation for sequentially acquired high-resolution images taken at equal time intervals. The model is helpful in tracking growth in plant development. The system was first tested with pilot experiments on baby bok choy but can be used for most crop types.

Penn State's research group, led by agricultural and biological engineers, has been practicing precision agriculture for more than a decade. They have their robotic crops on hand for harvest, pruning trees, green fruit thinning, pollination, orchard heating, spraying pesticides, and irrigation. Their machine vision system is founded on previous research in these fields.

During the trial, the machine vision system successfully separated individual baby bok choy plants in a soilless growth medium. Image analysis was applied consistently to regulate growing leaves' covering area, and precise growth was guaranteed. The recursive model showed stable operation during the growth period of the crop.

The research forms part of the umbrella federal projects, "Advancing the Sustainability of Indoor Urban Agricultural Systems," with the goal of increasing controlled environment agriculture's efficiency and sustainability. Based on the employment of AI and IoT technologies, the system tends to monitor automatically the crops' health, predict plant growth, and regulate environmental parameters such as radiation, temperature, and humidity. Through its data-dependent operation, it eliminates inefficiencies and makes greenhouse farming level.

Its potential scope can make even more subtle specialty crop quality and diet possible with the better cultivation through the accurate data.The study was a collaborative work among plant researchers and agricultural engineers. Researchers are Long He, associate professor of agricultural and biological engineering; Francesco Di Gioia, associate professor of vegetable crop science; Chenchen Kang, now a postdoctoral scholar; Xinyang Mu, postdoctoral scholar at Michigan State University; and Aline Novaski Seffrin, graduate student in plant science.The research was sponsored by the U.S. Department of Agriculture's National Institute of Food and Agriculture and the Pennsylvania Department of Agriculture.

Source: Penn State