In recent decades, the AECO (Architecture, Engineering, Construction and Operations) sector has undergone a significant digital transformation, driven by increasingly complex projects and performance and sustainability requirements. Although technological innovation began over fifty years ago, it is only in the last twenty years that advanced digital tools have accelerated change. The introduction of Building Information Modelling (BIM) has improved the management of the entire life cycle of a project, reducing costs and rework and optimising information flows. Despite this progress, there remains a gap between the use of BIM in the design phase and its application on site, where it could offer greater benefits in terms of quality control and management of operational variables. Managing large amounts of data, documents and activities is complex, mainly due to constantly changing operating conditions. Bridging this gap therefore requires constant and consistent updating of the information model, which follows the evolution of activities in the field in real time. In this context, this thesis proposes the development of an innovative system based on Computer Vision techniques, capable of automatically aligning images acquired on site with digital models, in order to update the scene in relation to the actual evolution of the operating scenario. This contributes to the construction of a digital scene composed of and enriched with heterogeneous data appropriately contextualised and synchronised with the actual progress of the work, in order to accurately and continuously document the evolution of the construction site. To make this system possible, a methodology for estimating image placement has been developed, based on direct comparison between images taken on site, without the need for physical markers, GPS or Wi-Fi networks. Testing of the system, carried out both in a controlled environment (laboratory) and on a real construction site, confirmed the effectiveness and reliability of the methodology even in complex and variable operating conditions. The system has made it possible to obtain a digital model that is constantly updated and enriched with data and information consistent with the actual status of the construction site, thus facilitating seamless integration between the physical environment and the virtual model. This represents a significant step towards the digitisation of work execution control and construction site management processes, outlining a concrete path for digital innovation in the construction sector.
Negli ultimi decenni, il settore AECO (Architecture, Engineering, Construction and Operations) ha affrontato una significativa trasformazione digitale, spinta da progetti sempre più complessi e da esigenze di prestazione e sostenibilità. Sebbene l’innovazione tecnologica sia iniziata oltre cinquant’anni fa, è solo negli ultimi vent’anni che strumenti digitali avanzati hanno accelerato il cambiamento. L’introduzione del Building Information Modeling (BIM) ha migliorato la gestione dell’intero ciclo di vita dell’opera, riducendo costi e rilavorazioni e ottimizzando i flussi informativi. Nonostante i progressi, rimane però un divario tra l’uso della modellazione BIM in fase di progettazione e la sua applicazione in cantiere, dove potrebbe offrire maggiori vantaggi in termini di controllo qualità e gestione delle variabili operative. La gestione di una grande quantità di dati, documenti e attività risulta complessa, soprattutto a causa del continuo cambiamento delle condizioni operative. Colmare questo divario richiede, quindi, un aggiornamento costante e coerente del modello informativo, che segua in tempo reale l’evoluzione delle attività sul campo. In questo contesto, la presente tesi propone lo sviluppo di un sistema innovativo basato su tecniche di Computer Vision, capace di allineare automaticamente le immagini acquisite in cantiere ai modelli digitali, al fine aggiornarne la scena, in relazione all’evoluzione reale dello scenario operativo. Questo contribuisce, alla costruzione di una scena digitale composta e arricchita di dati eterogenei opportunamente contestualizzati e sincronizzati con l’effettivo avanzamento dei lavori, al fine di documentare in modo accurato e continuo l’evoluzione del cantiere. Per rendere possibile questo sistema, è stata sviluppata una metodologia per la stima della posa delle immagini, basata sul confronto diretto tra le immagini scattate in cantiere, senza la necessità di marcatori fisici, GPS o reti Wi-Fi. La sperimentazione del sistema, svolta sia in ambiente statico (laboratorio) sia in un cantiere reale, ha confermato l’efficacia e l’affidabilità della metodologia anche in condizioni operative complesse e variabili. Il sistema ha permesso di ottenere un modello digitale costantemente aggiornato e arricchito con dati e informazioni coerenti con lo stato effettivo del cantiere, facilitando così un’integrazione fluida tra ambiente fisico e modello virtuale. Questo rappresenta un passo significativo verso la digitalizzazione dei processi di controllo dell’esecuzione dei lavori e della gestione del cantiere, delineando un percorso concreto per l’innovazione digitale nel settore delle costruzioni.
SISTEMA DI DOCUMENTAZIONE SEMANTICA DI MODELLI DIGITALI IN FASE DI ESECUZIONE SUPPORTATO DA COMPUTER VISION
GARDONI, CHIARA
2024/2025
Abstract
In recent decades, the AECO (Architecture, Engineering, Construction and Operations) sector has undergone a significant digital transformation, driven by increasingly complex projects and performance and sustainability requirements. Although technological innovation began over fifty years ago, it is only in the last twenty years that advanced digital tools have accelerated change. The introduction of Building Information Modelling (BIM) has improved the management of the entire life cycle of a project, reducing costs and rework and optimising information flows. Despite this progress, there remains a gap between the use of BIM in the design phase and its application on site, where it could offer greater benefits in terms of quality control and management of operational variables. Managing large amounts of data, documents and activities is complex, mainly due to constantly changing operating conditions. Bridging this gap therefore requires constant and consistent updating of the information model, which follows the evolution of activities in the field in real time. In this context, this thesis proposes the development of an innovative system based on Computer Vision techniques, capable of automatically aligning images acquired on site with digital models, in order to update the scene in relation to the actual evolution of the operating scenario. This contributes to the construction of a digital scene composed of and enriched with heterogeneous data appropriately contextualised and synchronised with the actual progress of the work, in order to accurately and continuously document the evolution of the construction site. To make this system possible, a methodology for estimating image placement has been developed, based on direct comparison between images taken on site, without the need for physical markers, GPS or Wi-Fi networks. Testing of the system, carried out both in a controlled environment (laboratory) and on a real construction site, confirmed the effectiveness and reliability of the methodology even in complex and variable operating conditions. The system has made it possible to obtain a digital model that is constantly updated and enriched with data and information consistent with the actual status of the construction site, thus facilitating seamless integration between the physical environment and the virtual model. This represents a significant step towards the digitisation of work execution control and construction site management processes, outlining a concrete path for digital innovation in the construction sector.| File | Dimensione | Formato | |
|---|---|---|---|
|
TESI GARDONI CHIARA S1098717.pdf
embargo fino al 18/04/2027
Dimensione
4.68 MB
Formato
Adobe PDF
|
4.68 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.12075/23575