Summary
Europe is home to about 400 UNESCO world heritage sites and has a growing tourism industry employing many people directly and indirectly. Hence, it is of concern to ensure the cultural heritage (CH) buildings are inspected properly and correct damage diagnosis is performed. Incorrect damage diagnosis will lead to loss of cultural value and may lead to the closing of the monument, thus affecting society in general and the livelihood of people associated with it. The proposed MLCULT project is motivated by the need to perform damage diagnosis of CH using image-based machine learning (ML) techniques, thus helping to preserve them. The popularity of ML approaches and deep learning algorithms has increased considerably over the last two decades. Computer-vision-based damage detection employing convolutional neural networks will be integrated with laser scan data, nondestructive testing, and other condition assessment data to provide a better estimate of existing areas of damage. The model will be trained from the database of earthquake-damaged CH collected by the host institutions UMinho and Polimi. Several typologies of damage indicators will be identified and quantified, due to weathering, moisture ingress, algae growth, and efflorescence. The project will be supervised by Prof. Lourenço at the University of Minho, Portugal, who is an international expert on CH, and Prof. Luigi Barazzetti at Politecnico di Milano, Italy whose has expertise in computer-vision, drone and image-based damage detection. Finally, a prototype inspection system (first of its kind in CH field) using drones-based real-time damage detection will be demonstrated, specifically for CH damage pathologies. The proposed method can help in identifying structural anomalies in CH that must be urgently repaired and thus can be used in high-quality civil infrastructure monitoring systems. This method would also enable fast screening of CH buildings after a disaster such as an earthquake.
Unfold all
/
Fold all
More information & hyperlinks
| Web resources: | https://cordis.europa.eu/project/id/101063722 |
| Start date: | 01-07-2022 |
| End date: | 30-11-2024 |
| Total budget - Public funding: | - 172 618,00 Euro |
Cordis data
Original description
Europe is home to about 400 UNESCO world heritage sites and has a growing tourism industry employing many people directly and indirectly. Hence, it is of concern to ensure the cultural heritage (CH) buildings are inspected properly and correct damage diagnosis is performed. Incorrect damage diagnosis will lead to loss of cultural value and may lead to the closing of the monument, thus affecting society in general and the livelihood of people associated with it. The proposed MLCULT project is motivated by the need to perform damage diagnosis of CH using image-based machine learning (ML) techniques, thus helping to preserve them. The popularity of ML approaches and deep learning algorithms has increased considerably over the last two decades. Computer-vision-based damage detection employing convolutional neural networks will be integrated with laser scan data, nondestructive testing, and other condition assessment data to provide a better estimate of existing areas of damage. The model will be trained from the database of earthquake-damaged CH collected by the host institutions UMinho and Polimi. Several typologies of damage indicators will be identified and quantified, due to weathering, moisture ingress, algae growth, and efflorescence. The project will be supervised by Prof. Lourenço at the University of Minho, Portugal, who is an international expert on CH, and Prof. Luigi Barazzetti at Politecnico di Milano, Italy whose has expertise in computer-vision, drone and image-based damage detection. Finally, a prototype inspection system (first of its kind in CH field) using drones-based real-time damage detection will be demonstrated, specifically for CH damage pathologies. The proposed method can help in identifying structural anomalies in CH that must be urgently repaired and thus can be used in high-quality civil infrastructure monitoring systems. This method would also enable fast screening of CH buildings after a disaster such as an earthquake.Status
SIGNEDCall topic
HORIZON-MSCA-2021-PF-01-01Update Date
09-02-2023
Images
No images available.
Geographical location(s)
