Indoor air quality in modern buildings is increasingly difficult to maintain without high energy costs, and while vertical green walls offer a natural solution, their inconsistent performance and complex maintenance have limited widespread use. VertINGreen, developed by researchers, solves this by using AI, remote sensing, and plant data to both predict how green walls will perform before installation and monitor their health in real time, making them a reliable, efficient, and scalable tool for improving air quality and reducing energy consumption.
Step into a modern office tower or hospital, and the air you breathe is often carefully engineered, filtered, circulated, and cooled at a high energy cost. Now imagine those same spaces quietly breathing on their own, supported by living walls of plants that not only beautify interiors, but actively clean the air and reduce energy use.
At the Hebrew University of Jerusalem, researchers have taken a major step toward that vision.
In a new study published in Indoor Air , Yehuda Yungstein and Dr. David Helman introduce VertINGreen , an innovative web-based platform that transforms vertical green walls from decorative features into intelligent, responsive environmental systems.
Vertical green walls have long captured the imagination of architects and designers. But behind their lush appearance lies a challenge: their performance is unpredictable. Some thrive and improve air quality; others struggle, requiring constant maintenance and offering little measurable benefit.
“Green walls have enormous potential,” the researchers explain, “but until now, we lacked the tools to truly understand and manage how they function indoors.”
VertINGreen changes that by bringing together remote sensing technology and machine learning, turning plant-covered walls into data-rich systems that can be planned with precision and monitored in real time.
The story of VertINGreen begins with nearly 2,000 meticulous measurements of how common indoor plants “breathe”, how they absorb carbon dioxide and release water under different conditions.
From this, the team built powerful predictive models capable of forecasting how a green wall will perform before it is even installed. The system can estimate:
The result is a planning tool that replaces guesswork with confidence.
“For the first time, designers can ask: What will this wall actually do for my building? and get a reliable answer,” says the team.
But VertINGreen doesn’t stop at planning, rather it continues the story long after installation.
Using hyperspectral imaging, the platform can “see” beyond visible light, detecting subtle physiological changes in plants. Combined with machine learning, this allows the system to:
In practical terms, this means fewer surprises, lower maintenance costs, and healthier, longer-lasting installations.
Remarkably, the system achieves this using just a handful of spectral bands, making it accessible even with relatively low-cost imaging equipment.
VertINGreen represents more than a technological innovation, it signals a shift in how buildings are conceived.
Instead of relying solely on energy-intensive systems, indoor environments can increasingly integrate living, adaptive components that work in harmony with technology.
By combining accurate planning with early-warning monitoring, VertINGreen offers a complete framework for:
And perhaps most importantly, it makes these benefits practical and scalable .
For Yungstein and Helman, the goal is clear: to bridge the gap between scientific understanding and real-world application.
“VertINGreen allows us to move from inspiration to implementation,” they say. “It gives architects, engineers, and building managers the tools they need to trust and fully utilize nature inside buildings.”
As cities continue to grow upward and inward, innovations like VertINGreen suggest a future where walls do more than divide space—they help sustain it.
Indoor Air
Experimental study
Not applicable
VertINGreen: A Practical Application for Planning and Monitoring Indoor Vertical Green Living Walls Based on Remote Sensing and Machine Learning Models
28-Jan-2026