Generic article

Why autonomous greenhouses?

Greenhouse production systems are ideal to protect crops against external influences (heat, snow, heavy rainfall, droughts, pests, etc.) and to control the production of healthy and tasty products. Greenhouses play an important role in all parts of the world in these times of climate change. They are very efficient in use of inputs such as water, fertilizer, and land, and very high in productivity.

Navigating complex greenhouse production systems

A greenhouse production system is complex. Many different factors influence vegetable growth (light, temperature, humidity, CO₂, water, nutrients etc.), and resource usage (energy, water etc.), as well as pests can disturb crop growth. Crop growth can also be influenced by various crop management strategies (plant density, fruit pruning and harvest strategies, fertigation, etc.). A grower must decide on the right setpoints for all parameters at every moment in time. He must deal with multiple growth parameter interactions and with multiple greenhouse actuators that affect the desired control (ventilation, heating, cooling, shading, artificial lighting, crop management actions etc.) and he must scout for pests and act accordingly by releasing predators or apply precision spraying.

Well-educated and experienced growers can oversee most aspects of such a system, but these growers are scarce worldwide. To feed a growing number of people in the future with a healthy, vitamin-rich diet, we need to develop new and fully automated production systems with low carbon footprints which can be operated by less experienced staff, too.

AI and computer vision could help

We assume that artificial intelligence, well-chosen machine learning and computer vision algorithms can optimize such a complex greenhouse crop production system. They also potentially give us better understanding of crop growth and interaction of all parameters, as well as new insights into sensor development and interpretation of sensor signals that can boost productivity while minimizing resource use.

This project is an important step to help growers with computers to better feed the world with healthy products and make optimum use of inputs like energy for a carbon zero society. In the fourth Autonomous Greenhouse Challenge we will use autonomous sensing and control to bring AI driven greenhouse horticulture closer to practical solutions. Like in previous editions we challenge teams to develop AI algorithms to make smart decisions on productivity and resource use to optimize resource use efficiency.”