While this solution isn’t new, progress is slow due to the high cost and specialised manpower required for maintaining such structures. But researchers are testing a promising alternative that employs drones and AI.
In a recent study conducted by researchers at the University of Rostock, these technologies were used to collect information about the condition of plants on roofs and facades and create their digital twins. A dataset of over 190 images of green and non-green surfaces was tested using a three-part setup:
- Drones for collecting real-time image data
- A digital twin for creating an interactive virtual representation of the building
- A multi-modal foundation model (MMFM), such as GPT-4, for analysing the images, identifying plant health issues, and producing care instructions.
The outcome of the study is promising: the AI model could accurately distinguish between vegetated and non-vegetated areas, identify plant species (in most cases), and detect visible signs of stress in plants, such as browning or damage. However, the study also exposed multiple areas for improvement, as there were inaccuracies related to identifying the exact species of plants or diagnosing plant damage in about 30% of test cases.
In addition to providing important information required for the maintenance of green roofs and facades, this setup offers a very efficient alternative for reaching hard-to-access areas compared to the traditional method of building costly and potentially dangerous scaffolds. Furthermore, the combination of drones and AI could also be used for automating such maintenance tasks as watering or pruning.
As climate pressures intensify, technologies could become the new standard for maintaining green infrastructure in our cities more efficiently and sustainably. This is also an important step in addressing the shortage and limited availability of qualified experts for plant care in urban areas with tricky access that requires special equipment and skills.
Reference:
Krause, N. S. & Ploennigs, J. Digital twins for green roofs and facade inspection using drones and multi-modal foundation models, E3S web of conferences, vol. 608, p. 05001, Jan. (2025). https://doi.org/10.1051/e3sconf/202560805001
Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or European Education and Culture Executive Agency (EACEA). Neither the European Union nor the granting authority can be held responsible for them.
