● RECONSTRUCT at AIVELA

On 26 June, RECONSTRUCT project partner UNIVPM organised a Special Session as part of the XVI International AIVELA Conference in Ancona.

Entitled ‘VISION SYSTEMS AND IMAGING TECHNIQUES IN THE CONSTRUCTION INDUSTRY: MEASUREMENT STRATEGIES TO SUPPORT SUSTAINABILITY AND CIRCULARITY‘, it was led by Gloria Cosoli (eCampus University), Giovanni Salerno (UNIVPM), Maria Teresa Calcagni (UNIVPM) and Gian Marco Revel (UNIVPM).

Prof. Revel chaired the special session, which explored the use of imaging techniques and vision systems to support sustainability and circularity in the construction sector with a focus on the valorisation of demolition waste. The aim was to promote innovative measurement strategies based on artificial intelligence to encourage the reuse and recycling of construction materials.

Three presentations were given as part of the RECONSTRUCT project:

• ‘Smart waste monitoring: an approach based on vision systems and AI for CDW management‘ (G. Salerno, M.T. Calcagni, G. Cosoli, L. Violini and G.M. Revel), presented by Dr Salerno (UNIVPM). This presentation demonstrated how a combination of a vision system and an AI-based algorithm can be used to identify and classify types of CDW waste material in real time in a real-world scenario. This scenario was based on the Sorigué waste management site in Spain, and the results were highly satisfactory, supporting the recycling processes from a sustainability point of view.
• The second presentation, titled ‘Validation of AI Tools for Construction and Demolition Waste Monitoring‘ and authored by H. de Melo Ribeiro, M.A.A. Abbas, A. Ayodeji, E. El Masri, G. Cosoli, G. Salerno, M.T. Calcagni and G.M. Revel, was presented by Dr de Melo Ribeiro (BUL). This presentation showcased the AI-based tools developed and validated for the detection, classification, and quantification of CDW. These tools were validated in three scenarios at two demonstration sites (COMSA and Sorigué). Images and videos were used to develop the method and train the AI models, achieving high accuracy.
• ‘Classification of Hyperspectral Data from Construction and Demolition Waste (CDW) Using the Spectral Angle Mapper (SAM) Algorithm‘ (authored by M. T. Calcagni, G. Salerno, G. Cosoli and G. M. Revel), presented by Dr Calcagni (UNIVPM). The presentation showed the Spectral Angle Mapper method for classifying selected types of CDW. This research demonstrates satisfactory results for classification and could provide a valuable foundation for future material reuse and emission reduction initiatives.