Process of converting laser scans into BIM models for energy simulation purposesAbstract
This study aims to develop a systematized workflow for converting laser scans, captured with mobile devices equipped with a LiDAR (Light Detection and Ranging) sensor, into BIM models compatible with energy simulations in Autodesk Revit. The methodology employs the pCon-Scan application to generate three-dimensional models, which are subsequently interpreted and converted into parametric geometries in Revit using Python scripts developed within Dynamo. The process was developed through iterative testing, initially applied to isolated environments and later extended to an entire building. The results indicated greater accuracy and consistency in single-environment models, while simultaneous scans of multiple spaces revealed certain geometric limitations. Nevertheless, the workflow demonstrated strong potential for automation, reducing manual steps and ensuring direct compatibility with Revit’s “Systems Analysis” tool. The process is currently in a refinement phase, seeking optimizations that enhance its efficiency, accessibility, and integration for the energy labeling of existing buildings.
Keywords
Laser scanning; BIM modeling; LiDAR; Energy simulation; Automation
Resumo
Este estudo tem por objetivo construir um fluxo de trabalho sistematizado para a conversão de escaneamentos a laser, capturados por dispositivos móveis equipados com sensor LiDAR (Light Detection and Ranging), em modelos BIM compatíveis com simulações energéticas no Autodesk Revit. A metodologia emprega o aplicativo pCon-Scan para gerar modelos tridimensionais, que são posteriormente interpretados e convertidos em geometrias paramétricas no Revit por meio de scripts em Python desenvolvidos no Dynamo. O processo foi desenvolvido a partir de testes iterativos, aplicados inicialmente em ambientes isolados e, posteriormente, em um edifício completo. Os resultados indicaram maior precisão e consistência nos modelos de ambientes únicos, enquanto as varreduras simultâneas de múltiplos espaços apresentaram certas limitações geométricas. Ainda assim, o fluxo demonstrou elevado potencial de automação, reduzindo etapas manuais e garantindo compatibilidade direta com a ferramenta “Systems Analysis” do Revit. O processo encontra-se atualmente em fase de aprimoramento, buscando otimizações que aumentem sua eficiência, acessibilidade e integração para a etiquetagem energética de edificações existentes.
Palavras-chave
Escaneamento a laser; Modelagem BIM; LiDAR; Simulação energética; Automação
1 Introduction
The growing demand for more sustainable buildings has intensified both global and national interest in mechanisms capable of assessing and improving the thermal and energy performance of buildings, such as energy labeling systems. In Brazil, this process is regulated by the Programa Brasileiro de Etiquetagem de Edificações (PBE Edifica), which aims not only to inform consumers about a building’s efficiency level but also to encourage designers, builders, and owners to adopt more energy-efficient solutions throughout all stages of the building’s life cycle.
However, the challenges related to the practical application of energy labeling in existing buildings extend far beyond the certification process itself. Before an inspection body (Organismo de Inspeção Acreditado – OIA) can perform a proper efficiency evaluation, it is necessary to obtain a digital representation of the building that accurately reflects its geometric and constructive characteristics. This preliminary stage – which encompasses the survey, modeling, and parameterization of architectural elements – remains largely dependent on manual and time-intensive procedures. Such methods are not only slow and costly but also demand the involvement of additional professionals dedicated exclusively to creating the digital model, thereby increasing expenses and delaying the issuance of the energy label.
In addition, previous studies have highlighted differences between simplified approaches and simulation-based methods for energy performance assessment, emphasizing the higher level of geometric consistency and modeling detail required by simulation workflows (Pimentel et al., 2021). In this sense, simulation-based approaches impose more stringent requirements on the digital model, particularly regarding the coherence of architectural elements and the integrity of spatial definitions.
The limited number of accredited inspection bodies in Brazil further exacerbates this situation, restricting the country’s overall capacity to meet the growing demand for building energy efficiency certifications. In this context, the development of integrated and automated solutions becomes essential to overcome these structural barriers. The approach proposed in this research seeks to incorporate the digital modeling stage into the broader energy labeling workflow, streamlining the process from data acquisition to energy simulation. By enabling the same professional team to conduct scanning, model generation, and analysis within a unified environment, this method aims to make the process faster, more economical, and less fragmented. Consequently, the proposed workflow adopts energy simulation not as an end in itself, but as a validation strategy for the generated models within a Building Information Modelling (BIM) framework, allowing the method to be tested under more rigorous analytical conditions. This choice reinforces its potential applicability in performance-oriented BIM workflows, while maintaining technical reliability and reducing the burden on both professionals and end users.
Within this scenario of limited inspection infrastructure, technologies capable of automating geometric data collection emerge as key enablers for expediting the entire labeling workflow. The digitalization of existing buildings through laser scanning has become an essential resource in architecture and engineering, particularly in renovation, restoration, and interior design projects (Autodesk, 2024; Borkowski; Kubrat, 2024; Sadeghineko; Lawani; Tong, 2024; Wang; Guo; Kim, 2019). Among the available technologies, Light Detection and Ranging (LiDAR) – integrated, for instance, into devices such as the iPhone 13 Pro – offers a practical and portable means of capturing three-dimensional spatial data directly on-site (Apple, 2024). LiDAR operates by emitting laser pulses and measuring the time taken for each pulse to return after striking a surface, thereby generating a dense point cloud that accurately represents the geometry and spatial configuration of the scanned environment.
Currently, only Apple® mobile devices – specifically the Pro versions of iPhones and iPads released from 2020 onwards – feature built-in LiDAR sensors (Cuperschmid; Neves de Oliveira; Froner, 2024). This illustrates a growing technological trend: the integration of advanced sensing and spatial computing capabilities into everyday portable devices. The accessibility, practicality, and intuitive use of such equipment have the potential to democratize spatial data acquisition, making digital surveying more attainable for architects, engineers, and construction professionals, even without access to high-cost specialized equipment.
Nonetheless, despite these remarkable advances, the geometries generated by such applications still cannot be directly imported or converted into native BIM elements within modeling software such as Revit, developed by Autodesk. This limitation significantly restricts the ability to edit materials, assign thermal and physical properties, or integrate analytical parameters required for accurate energy performance simulations.
Considering the recent technological advancements that allow energy simulations to be conducted directly within Revit through the “Systems Analysis” tool (Rodrigues, 2023), this study builds upon the existing body of literature on the scan-to-BIM process – that is, the transformation of scanned spatial data into BIM models, specifically Revit models, suitable for simulation and analysis. In this context, this study aims to develop a systematized workflow for converting laser scans, captured with mobile devices equipped with a LiDAR (Light Detection and Ranging) sensor, into BIM models compatible with energy simulations in Revit.
2 Theoretical framework
The integration between three-dimensional surveying technologies and BIM modeling environments has become an increasingly recurrent and relevant topic within international research, reflecting the ongoing evolution of digital workflows in architecture, engineering, and construction. For instance, Macher, Landes and Grussenmeyer (2017) proposed a semi-automated reconstruction method for existing buildings based on the processing of point cloud data. Their methodology consisted of successive segmentations of the point cloud, followed by the identification of planar surfaces corresponding to architectural components such as walls, ceilings, and floors. Once these surfaces were defined, the structural elements were described in .obj format and reconstructed according to the extracted geometric data, taking into account diverse configurations and spatial relationships. Although this approach substantially reduced the overall modeling time when compared to traditional manual methods, it still required significant human intervention to guarantee model accuracy and geometric consistency – highlighting the persistent challenges in achieving a fully automated and reliable scan-to-BIM workflow.
In a more recent contribution, Zeng et al. (2023) developed the Auto-Scan-To-BIM system, specifically designed to automate the conversion of point cloud data into BIM-compatible models. Their system comprises three main modules: the first performs enhanced plane segmentation through edge detection and corner recalibration; the second applies heuristic rules grounded in construction domain knowledge to identify architectural elements – such as walls, floors, ceilings, windows, and doors – and to generate their corresponding BIM components; while the third module exports the final model in the Industry Foundation Classes (IFC) format, ensuring interoperability with a wide range of BIM platforms and analytical tools. The results obtained by Zeng et al. (2023) demonstrated high accuracy in the automatic identification of architectural elements and in the generation of detailed and semantically rich BIM models, contributing to practical applications in quality control, construction planning, and building energy performance evaluation. Nevertheless, the authors acknowledged persistent limitations, including difficulties in representing complex geometries, irregular surfaces, and in precisely determining component thicknesses – aspects that continue to present opportunities for further refinement and technological advancement.
According to Gleń and Krupa (2019), the automation of architectural surveys using LiDAR-generated point clouds offers considerable productivity gains, far surpassing the efficiency of conventional documentation methods based on manual measurements and CAD-based redrawing. In their comparative study, replacing the traditional documentation process with a digital workflow resulted in a 52% increase in overall productivity, clearly illustrating the positive impact of technological integration in optimizing architectural documentation, modeling, and information management. These findings establish a robust reference framework for the present research, providing a measurable benchmark of efficiency and supporting the argument for automation within scan-to-BIM processes.
Recent literature investigating the coupling of reality-capture technologies with BIM environments underscores the critical role of data interoperability and format standardization in enabling seamless workflows from scanning to modeling. For instance, Borkowski and Kubrat (2024) emphasize that once a point cloud is acquired, the choice of file format depends not only on the scanning device and the processing software used but also on the target BIM platform and on the potential representation of the model within that environment. Their analysis points out that commonly adopted formats, such as RCP (ReCap Project), for Revit, and E57 (ASTM E57 3D Imaging Data File Format), XYZ, or LAS (LASer file format) for other software – each possess distinct capacities and limitations regarding geometric fidelity, semantic data retention, and downstream usability. Furthermore, the authors highlight that the final step of the scan-to-BIM process typically involves exporting the model in an open standard such as IFC, in order to maximize interoperability across different platforms, even though complete and lossless data exchange is not yet guaranteed. This observation reveals that, while digital capture of spatial geometry is increasingly mature, the effective translation of that geometry into semantically structured BIM elements remains highly dependent on data format compatibility, metadata richness, and software recognition of native components.
As emphasized by Borkowski and Kubrat (2024), the degree of interoperability between scanning outputs and BIM environments is strongly influenced by the adopted export formats and by how effectively modeling software can interpret and translate the resulting geometries. In this regard, workflows based on point cloud data frequently present considerable challenges, as they often require multiple processing steps, specialized technical expertise, and a substantial amount of manual adjustment before the information can be efficiently integrated into BIM systems. These complexities tend to slow down the modeling process and limit the practical adoption of digital scanning technologies within architectural and energy performance analysis workflows.
To simplify such procedures, several mobile applications have recently introduced functionalities capable of automatically translating spatial data captured through LiDAR sensors into simplified architectural representations, identifying elements such as walls, doors, windows, and openings. These innovations provide a more direct, accessible, and user-friendly means of generating digital models of existing environments. This technological evolution has been made possible due to Apple’s RoomPlan API, introduced in 2022, which enables the automatic creation of three-dimensional indoor models by combining LiDAR-based scanning with machine learning algorithms and augmented reality frameworks (Apple, 2022). Rather than relying exclusively on raw point clouds, these systems interpret captured data in real time, reconstructing the environment using geometric primitives and semantic recognition of architectural components. The outcome is a structured, lightweight, and organized 3D model that enhances efficiency and accessibility compared to traditional scanning and post-processing techniques.
3 Method
This study adopts the Design Science Research (DSR) method, which focuses on creating artifacts – such as processes, systems, and methods – to solve practical problems through innovation (Dresch; Lacerda; Antunes, 2014). As highlighted by Hevner, Park and March (2019), DSR is particularly suitable for generating both theoretical and practical contributions in the field of digital innovation.
In this research, a workflow was developed that integrates data capture, modeling, and energy simulation into a continuous and systematized process. The study investigated methods for 3D scanning of existing buildings and strategies for seamless integration with BIM, aiming to reduce manual steps and produce models directly compatible with simulation tools.
Following Bunge’s (1980) model, the development process adopted an iterative approach encompassing problem identification, artifact design, evaluation, and refinement, emphasizing practical application and continuous improvement of the proposed technological solution. The DSR approach is structured into three phases: knowledge exploration, focused on understanding the underlying problem; process construction, which involves developing the proposed solution based on the requirements identified during the exploration phase; and method evaluation, comprising the testing and validation of the effectiveness of the developed solution.
3.1 Knowledge exploration
This initial stage is related to what Lacerda et al. (2013) define as the awareness phase in the DSR process, in which the problem to be addressed is identified and delimited, including an understanding of its external context and the characterization of desirable solutions.
Accordingly, the digital tools to be further investigated in this study were defined, considering their potentialities and limitations in relation to the proposed objective. These tools, employed throughout the process, comprise scanning applications, modeling software, and energy efficiency simulation software.
Different LiDAR-supported scanning applications were selected for testing in order to analyze their practical applicability in capturing three-dimensional data of buildings, considering their functionalities, limitations, and potential for integration with BIM modeling. The tests were conducted consistently in the same location, using only applications that rely on the LiDAR sensor. Based on these tests, only applications that included the RoomPlan mode were selected for further comparison across additional criteria. After the scanning process, the three-dimensional data were exported in different formats. Tests were conducted to evaluate data compatibility and integrity when imported into Revit, with a focus on generating BIM models containing native wall, door, and window elements.
It is important to emphasize that these tests do not yet represent the process proposed as the solution to the research problem. Rather, they constitute a preliminary study aimed at analyzing the technological possibilities currently available on the market, with the objective of understanding their limitations, potentials, and implications for the generation of BIM models compatible with simulation processes.
3.2 Process construction
In Design Science Research (DSR), artifact construction may adopt different approaches, such as computational algorithms, graphical representations, or prototypes (Lacerda et al., 2013). In this study, the artifact is implemented as a computational and systematized workflow, designed to be functional, replicable, and applicable to a broader class of problems related to scan-to-BIM processes (Venable, 2006; Hevner; Park; March, 2004). The workflow was developed using Autodesk Revit 2025, with Dynamo integrated into the platform, and Python scripting executed within the Dynamo environment. Spatial data were acquired using the built-in LiDAR sensor of an iPhone 13 Pro.
During this phase, the focus shifted from exploring available market solutions to testing and applying alternatives that support the development of the proposed workflow. Based on the formats identified during the exploratory tests, three alternative strategies were considered for incorporating scanned data into Revit:
-
the use of pCon Planner as an intermediary platform for format conversion;
-
the direct import of IFC files, a widely adopted open standard for BIM interoperability (ISO, 2024); and
-
the development of routines in Dynamo aimed at generating native Revit elements from scan-derived data.
For the purposes of structuring the proposed artifact, the third strategy was selected as the reference approach, serving as the basis for organizing the workflow developed in this study. Based on this decision, the methodology was systematized into five sequential stages: Data Acquisition, Preliminary Geometric Processing, Geometry Interpretation, Parametric Modelling, and Energy Simulation. These stages define the analytical framework adopted in the Results section, where each step is examined in detail. The Data Acquisition stage includes controlled scanning tests involving specific geometric configurations – such as double-height spaces, level differences between rooms, curved walls, and concave volumes – selected to isolate potential geometric challenges and assess their impact on the applicability of the proposed method. The Preliminary Geometric Processing stage describes how the scanning application generates a simplified geometric representation of the environment based on the RoomPlan framework. The Geometry Interpretation stage details the structure of the exported scan files, the parameters extracted for each architectural element, and the Python-based procedures used to interpret spatial data, including coordinate transformations and rotation handling. The Parametric Modelling stage explains the visual programming logic implemented in Dynamo, in which the interpreted data are mapped to user-selected Revit families to generate native BIM elements. Finally, the Energy Simulation stage evaluates the compatibility of the resulting model with Revit’s Systems Analysis tool, using default simulation parameters, with the objective of verifying whether the generated BIM model can be directly employed for energy analysis without additional manual adjustments.
3.3 Method evaluation
At this stage, the proposed workflow was evaluated through a comparative case study, in which the conventional survey and modeling method and the proposed LiDAR-based method were applied to the same environment. The objective of this evaluation was to assess the feasibility and effectiveness of the proposed process in comparison with the conventional approach, considering the ability to produce an equivalent BIM model and the overall efficiency of the workflow.
Based on these three phases, the proposed methodology effectively combines automated building digitization with a fully parameterized modeling and computational analysis process developed within the Revit environment. The central focus of this study is the optimization of the surveying and geometric modeling stages for existing buildings, taking advantage of LiDAR-equipped mobile devices and visual programming techniques to produce BIM-compatible models ready for subsequent analysis. To confirm the adequacy and compatibility of the generated models for energy simulation purposes, the workflow described by Rodrigues (2023) was employed exclusively in the analysis stage within Revit. Therefore, it is important to emphasize that the simulation itself does not constitute the main object of this research, but rather functions as a verification mechanism for validating the proposed process of geometric conversion and model generation. This approach significantly minimizes the need for extensive manual remodeling in external applications, optimizing both the time and the resources required for conducting energy performance assessments and certifications of existing buildings.
In accordance with the Design Science Research framework, the development process was conducted through iterative and practice-based experimentation. The proposed workflow was continuously refined through real-world testing, which involved scanning different environments using LiDAR-equipped mobile devices and analyzing the resulting data structures.
4 Results and discussion
This section presents and discusses the results obtained from the application of the proposed methodology, whose main outcome is a five-step systematized workflow for the digitization and preparation of existing buildings for energy analysis. The results are organized according to each stage of the workflow, from LiDAR-based data acquisition to model preparation for energy simulation. For each stage, the observed outcomes are presented alongside a discussion of their implications, limitations, and contributions to the overall effectiveness of the proposed process. This integrated approach enables a comprehensive assessment of how the methodology performs in practice and to what extent it supports the generation of BIM models suitable for computational energy analysis.
4.1 Data acquisition
Prior to conducting the scanning tests, different mobile applications with LiDAR support were analyzed in order to assess their data export capabilities and their potential integration with BIM workflows. Although several tools offer similar scanning functionalities, the ability to export structured and readable data proved to be a decisive criterion. Despite its relatively simple interface and feature set, the pCon.scan application demonstrated significant potential due to its capability to export scan results in a text-based format, enabling direct access to the underlying geometric descriptions through XML files. This characteristic was essential for the subsequent stages of data interpretation and parametric modeling.
Specific geometric challenges were examined, focusing on configurations that could potentially interfere with the scanning and modeling of interior environments when applying the proposed method. To this end, a series of tests was conducted to evaluate the extent to which these conditions affected the workflow, determining whether they constituted limiting factors or whether viable strategies could be adopted to allow the continued use of the method.
4.1.1 Ceilings with different heights
During the tests involving specific geometries, the analysis focused mainly on the behavior related to the generation of walls and openings, since horizontal elements such as floors and ceilings are not directly captured by the pCon.scan application. In the adopted workflow, the ceiling modeling in the BIM environment was based on the final height of the generated walls, which served as the only reference for determining the ceiling level. However, the scanning process did not recognize more complex geometries in the ceilings, being limited to the vertical surfaces that define the perimeter of the space. As a result, the generated model provides a simplified representation of the built reality, requiring additional visual verification, through photos, videos, or on-site inspection, to ensure that the modeled volumetry accurately reflects the existing conditions.
4.1.2 Walls junction
In spaces where the walls form concave angles (greater than 180°), specific geometric errors were identified in the models generated by the pCon.scan application. In such cases, the connection between walls is not properly represented in the 3D model, resulting in a visual distortion that resembles a chamfered pillar at the junction point. This issue can already be noticed in the app visualization and becomes more evident after conversion to Revit, where, through the interpretation of the XML code, the walls end up overlapping improperly.
Additionally, in orthogonal configurations (right angles), small connection errors were occasionally observed. Unlike in the concave cases, these errors resulted in small gaps between wall elements rather than overlaps. Despite these geometric inaccuracies, both situations can be easily corrected directly in Revit using the Trim command, which allows for quick manual adjustments to ensure wall continuity. Therefore, this is a point that could be improved in future updates of the application to enhance spatial fidelity.
4.1.3 Glass curtain walls
When using LiDAR-based mobile scanning, glass curtain walls proved problematic. When a glass surface extends from floor to ceiling, the pCon.scan application often fails to recognize it as either a wall or an opening, resulting in gaps in the generated model. Recognition only occurs when there is an opaque wall portion either above or below the glass surface, allowing the application to classify it as a window or glass door. Figure 1 illustrates an example of a scanning attempt in a glazed façade, where several trials were made but the element was not detected, highlighting the method’s limitations in contexts involving transparent materials.
4.1.4 Curved walls
Curved walls represent a significant challenge for scanning-based workflows relying on the RoomPlan framework. During the scanning process, curved surfaces are discretized into a sequence of vertices and subsequently simplified by the application’s internal algorithm into straight segments. As a result, the original curvature is lost and replaced by a faceted approximation of the geometry, as illustrated in Figure 2. This behavior directly affects the geometric fidelity of buildings containing curved architectural elements and requires subsequent manual correction or remodeling to preserve the original design intent. Similar limitations related to complex geometries and irregular surfaces have been reported by Zeng et al. (2023), reinforcing that such configurations remain a critical challenge in scan-to-BIM processes and still represent an open field for methodological refinement.
4.1.5 Multiple rooms in a single scan
This test evaluated the performance of the proposed method in situations involving spatial continuity and the connection between multiple rooms, assessing the ability of the system to maintain geometric consistency and data organization throughout an extended and interconnected scanning process. This type of configuration is particularly relevant for validating the applicability of the method in larger environments, where junction precision and spatial coherence are essential for model quality.
The tested environment contained eight simple rooms with relatively straightforward geometries. However, since the capture was performed as a continuous scan while moving from one room to another, the resulting model displayed inconsistencies and several geometric errors, as shown in Figure 3. More advanced applications, such as MagicPlan, already include a MultiRoom feature, which allows scanning multiple rooms separately while preserving their spatial relationships through a unified coordinate system—an ideal configuration for extended environments. However, despite offering more advanced scanning functionalities, these applications did not enable the level of BIM integration achieved with pCon.scan, particularly with respect to accessing and interpreting structured data required for the generation of native BIM elements. Until such functionality is implemented, the proposed workflow requires scanning each room individually and then manually rotating and positioning them correctly within Revit, which introduces an additional organizational effort.
4.2 Preliminary geometric processing
In the literature, the most common alternative to conventional manual modeling involves the use of point clouds extracted from 3D scans to reconstruct BIM models. However, these scan-to-BIM workflows are frequently associated with multiple processing stages, the need for specialized technical expertise, and a substantial amount of manual intervention before effective integration into BIM environments (Borkowski; Kubrat, 2024). In the present study, a different scenario is explored: instead of relying on raw point clouds, the workflow builds upon the structured geometric model generated by Apple’s RoomPlan framework, shifting the focus from geometric reconstruction to data interpretation.
The point cloud obtained during data acquisition is processed by the scanning application through algorithms based on Apple’s RoomPlan framework. This processing combines LiDAR depth data with device motion and image-based recognition to identify planar surfaces and infer basic architectural elements, such as walls, doors, windows, and room boundaries.
The result of this step is an automatically generated, simplified representation of the scanned environment, in which architectural components are recognized and classified according to predefined categories. Although this representation is sufficient for spatial organization and visualization, it does not provide direct control over parametric or analytical attributes required for BIM-based modeling and energy analysis. Consequently, this preliminary processing serves as a structuring layer that organizes the raw scan data and enables its export for subsequent stages of computational interpretation and parametric modeling.
4.3 Geometry interpretation
Autodesk Revit was selected as the modeling environment in this study due to its capacity to integrate geometric modeling and energy efficiency simulation within a single platform, thereby minimizing rework and reducing dependence on multiple software tools. This choice is aligned with the approach presented by Rodrigues (2023), who demonstrates the feasibility of using Revit’s Systems Analysis feature to generate analytical models for energy simulation. From the early stages of the research, this decision guided the structuring of the workflow toward enabling the direct incorporation of digitally captured geometry into a BIM environment compatible with performance analysis.
To enable the transformation of scan-derived data into BIM-compatible information, this stage focused on the interpretation of structured files generated during the scanning process. The export format of mobile scanning applications proved to be a determining factor for computational interpretation and subsequent model generation. Initial strategies explored in this research included the use of intermediary software for format conversion and the direct import of IFC files. However, as highlighted by Borkowski and Kubrat (2024), the choice of file format in scan-to-BIM workflows is strongly dependent on both the target BIM platform and the capacity of that platform to interpret and reconstruct native elements. Although IFC is widely adopted as an open interoperability standard, its use does not guarantee complete or lossless data exchange, often resulting in models with limited editability and reduced parametric control. These limitations were also observed during the exploratory tests conducted in this study, as Revit was unable to reliably recognize imported geometries as fully native BIM elements suitable for further parametrization and analysis.
A key advancement was the identification that the pCon.scan application allows the export of scan data in a text-based format, packaged in a file with the .eox extension. This file consists of a compressed set of XML (Extensible Markup Language) documents, including meta.xml, objects.xml, programs.xml, and structure.xml. Among these, the structure.xml file was identified as the most relevant, as it contains a detailed and hierarchical description of the scanned elements through structured instructions.
Rather than attempting to reinterpret already-formed three-dimensional geometry, the proposed approach focused on directly reading and interpreting this underlying code. To accomplish this, a Python script was developed within the Dynamo environment to parse the contents of the structure.xml file, extract geometric and parametric information, and organize it in a form suitable for subsequent BIM modeling.
Based on this premise, the developed Python script performs the conversion of scanned data by interpreting architectural object descriptions contained in the XML structure and translating them into geometric instructions. The xml.etree.ElementTree library was employed to parse the hierarchical organization of the XML files, providing the logical framework for extracting information required for geometric reconstruction.
For each object classified as a “wall”, parameters such as length, height, starting point, and rotation were extracted directly from the XML data. Rotations were represented using quaternions, defined by four components (w, x, y, z), which offer a mathematically robust method for describing three-dimensional orientation while avoiding singularities commonly associated with Euler angles.
To ensure compatibility with the modeling environment, all dimensional values were converted from meters to feet, in accordance with Dynamo’s internal unit system. Additionally, spatial coordinates were transformed to align with Dynamo’s coordinate reference, preserving the X-axis while swapping the Y and Z axes (X, Y, Z → X, Z, Y), thus ensuring correct orientation and spatial consistency within Revit.
The interpretation of object rotation required specific implementations for quaternion-based transformations, including the calculation of the quaternion conjugate and the application of quaternion multiplication to rotate points in three-dimensional space. Accurate handling of these operations was essential to preserve the spatial positioning and orientation defined during the scanning process.
4.4 Parametric modeling
Building upon the interpreted data obtained in the previous stage, this phase focused on the generation of native BIM elements within the Revit environment. Dynamo was employed as a visual and procedural programming interface, enabling the connection between the Python-based data interpretation routines and Revit’s modeling capabilities through the Revit API. Further implementation details are discussed in the author’s master’s dissertation (Bertoldi, 2025).
Based on the structured information extracted from the XML file, parametric routines were implemented to generate architectural elements such as walls, doors, and windows as native Revit objects. This approach allowed the assignment of geometric properties and basic material and thermal parameters, ensuring consistency with BIM standards and compatibility with subsequent analytical use.
After the creation of the parametric volumetry and the necessary adjustments to the constructed elements, the resulting BIM model reached the level of completeness required for computational analysis. At this point, the process transitioned from geometric modeling to analytical preparation, establishing the basis for energy simulation within the Revit environment.
To ensure consistency and reproducibility of the workflow, the modeling routine incorporated selection nodes that allow users to define specific Revit families for walls, doors, and windows. The Python script then associates the dimensional and positional information extracted from the scan data with the selected Revit families, enabling the automatic assignment of geometry, dimensions, and basic parametric attributes. This approach allows the method to be applied across different projects while maintaining control over the parametric and material characteristics of the generated BIM elements.
A preliminary test was carried out in a small-scale environment consisting of a door, a window, and an opening connecting to a corridor with four additional doors, as shown in the scan model in Figure 4(a). The generated elements still required further adjustments and interpretations to fully comply with the BIM workflow, particularly within Revit. However, through iterative testing and verification across different cases, the integration process was successfully validated, ensuring that the models maintained geometric accuracy.
(a) 3D model generated during scanning; (b) Overview of the Dynamo environment; (c) Enlarged view of the Dynamo environment, with the visual programming organized into stages for clearer reading of the workflow
Given its higher level of interoperability with Revit, pCon.scan proved to be the preferred tool for this workflow, even though it lacks some of the convenience features available in other applications. Despite its simplicity, the main objective was achieved: the automatic generation of a BIM-compatible model capable of supporting energy efficiency simulations. Figure 4(b) presents a zoomed-out view of the Dynamo environment, providing a general overview of the visual programming workspace and the connections derived from the XML file commands, which are translated into the Revit model. Figure 4(c) shows a zoomed-in view of the same environment, organized into distinct stages to facilitate the reading and interpretation of the visual programming.
The fidelity of the XML data conversion to Revit were visually evaluated by comparing the generated model with the original plans from several tests. The images (Figure 5) show that the created elements follow the positions defined in the XML file. Small discrepancies may occur due to variations in the processing of the rotation quaternions, but the impact on the overall modeling was minimal.
Comparison of original geometry from scans (a) and reproduction in Revit geometry from the implementation (b).
Despite the efficiency of mobile LiDAR-based scanning applications, certain geometric discontinuities and incomplete elements frequently appear in the exported data. The most common problems observed included gaps or overlaps at wall junctions, as well as the absence of structural components that are either not easily detectable by mobile sensors or too complex to be simplified into a single constructive element.
To address these issues and prepare the model for the subsequent energy analysis, the walls were manually joined, ensuring geometric continuity throughout the environment. Additionally, elements such as the floor slab, roof slab, and ceiling were created manually within Revit, since these components are not automatically generated by the scanning process yet. These additions are crucial for defining thermal zones and boundary conditions in the energy model, as they provide the necessary enclosures for the analytical geometry to be correctly interpreted by the simulation engine. With these adjustments completed, it became possible to proceed with the energy simulation directly in Revit.
4.5 Energy simulation
Once the scanned models were validated and adjusted to meet the necessary precision standards, the workflow could advance toward the next stage – the transformation of these architectural representations into analytical energy models, as illustrated in Figure 6(a). This transition marked an important step in verifying whether the geometries obtained from LiDAR-based mobile scanning could effectively support the simulation of building energy performance within BIM-integrated environments.
(a) Analytical model generated from the scanned model; (b) Energy analysis report generated within Revit software.
The process outlined by Rodrigues (2023) aims to transform the architectural model into an analytical energy model that complies with established standards for thermal performance evaluation. The energy simulation process involves a series of interdependent steps that must be carefully configured to ensure accuracy and consistency. Initially, the project’s geographic location and corresponding climate data are defined, establishing the environmental conditions that will influence the thermal calculations. Subsequently, the constructive properties of each element – such as the materials and thermal characteristics of walls, openings, slabs, and other building components – are assigned to the model. These parameters are essential to reproduce the real physical behavior of the materials and ensure that the simulation results accurately reflect the building’s actual performance. Next, the internal spaces are created and classified within the BIM model, followed by the configuration of the analytical parameters that control the energy simulation process. These settings ensure that all environments are correctly interpreted and connected by the simulation engine, allowing for coherent analysis of heat exchange, solar gains, and thermal comfort metrics across the model.
Once all of this information is properly entered and verified, the model is processed through Revit’s native energy analysis interface, which operates using the EnergyPlus simulation engine in conjunction with the OpenStudio platform. The results, shown in Figure 6(b), include data such as total energy consumption, heating and cooling loads, and thermal comfort indicators. This workflow demonstrates that, with minimal geometric corrections and proper parameterization, models derived from mobile scanning technologies can be effectively integrated into BIM-based energy simulation environments, thus expanding their applicability beyond documentation and into the realm of performance analysis. By leveraging these advanced analytical tools through a unified platform, the method preserves the precision and reliability of traditional simulations while reducing the complexity of data exchange between different software environments. This results in a streamlined, reproducible, and user-friendly workflow that supports the automated evaluation of a building’s energy efficiency without requiring additional export or reconfiguration steps.
4.6 Geometric tests as method evaluation
A comparative study was carried out to evaluate the efficiency of the proposed method in relation to the conventional approach for surveying and modeling existing environments. The objective of this preliminary assessment was not to establish definitive conclusions, but rather to provide an initial understanding of how the integration of mobile scanning could influence time and effort in the modeling workflow.
The conventional method relied on manual measurements using a measuring tape, complemented by hand sketches for recording the collected data. The process involved two people – one responsible for measuring and another for annotating and drafting the spaces. The recorded dimensions of the rooms and openings were then manually transferred into the modeling software, where the architectural model was reconstructed based on the field notes. The images in Figure 7 illustrate this procedure, from the physical survey to the final digital model generated in Revit.
In contrast, the proposed method employed the LiDAR sensor of an iPhone and a systematized data processing workflow. The entire process was carried out by a single person and completed in a significantly shorter time. For comparison purposes, both the conventional and the proposed methods aimed to achieve equivalent modeling outcomes, resulting in digital models that met the same requirements established for this study. Therefore, the total time measured for the proposed method included any necessary manual adjustments to the model, such as the alignment or merging of wall elements, ensuring a fair comparison between the two approaches. Figure 8 illustrates the stages of this process, from scanning the space to visualizing the modeled environment within Revit.
The comparison considered metrics such as the area of rooms, ceiling height, window area, number of people involved, and total time required for data acquisition and modeling. As summarized in Table 1, the two methods produced highly consistent geometric results.
In both analyzed environments, the total floor areas obtained through the proposed method coincided with those measured conventionally, showing no difference. Ceiling height values were also identical for one of the rooms and presented only a minor variation of 0.43% in the other. The largest deviation was observed in the window area, with a difference of 4.52%. From an operational standpoint, notable advantages were observed. The proposed method required only one person, while the conventional approach demanded two. The total execution time was also considerably reduced – by approximately 3.6 times, decreasing from 18 minutes to just 5 minutes.
Although these findings are limited to the specific case studied, they indicate that the proposed workflow can provide substantial efficiency gains without compromising accuracy. In this context, the method proved to be leaner, faster, and more practical, suggesting its potential for future applications in buildings with similar characteristics.
After refining the scanning-to-modeling workflow, the proposed process reached a level of consistency that allowed for the generation of accurate architectural models suitable for analytical use. The iterative tests carried out throughout the study demonstrated that, although certain manual interventions were still required, the systematized workflow ensured geometric fidelity and parametric control over the generated elements.
5 Conclusions
This study presented the development and evaluation of a systematized workflow for converting LiDAR-based mobile laser scanning data into BIM models compatible with energy simulation within the Revit environment. Aligned with the principles of Design Science Research, the proposed methodology integrates automated building digitization, parametric modeling, and computational analysis into a coherent process aimed at reducing manual modeling effort and improving workflow efficiency.
The main contribution of this research lies in demonstrating the feasibility of directly interpreting structured scan data to generate native BIM elements, enabling greater automation and consistency when compared to conventional scan-to-BIM approaches. By focusing on data interpretation rather than geometric reconstruction, the workflow establishes a robust link between digital surveying and BIM-based energy analysis, supporting applications related to energy labeling and performance assessment of existing buildings.
The results obtained through iterative testing and case study application indicate that the proposed process is functional and effective within the scope evaluated. However, the methodology is not presented as a finalized or commercial-ready solution. Certain limitations – particularly related to complex geometries and multi-space environments – highlight the need for further refinement and optimization of the automated procedures.
Future developments of this research include the consolidation of the workflow into a dedicated mobile application capable of integrating scanning, data interpretation, and model generation into a unified process, as well as the expansion of automation toward analytical model generation and cloud-based energy simulation. These advances are expected to further reduce manual intervention and strengthen the integration between digital surveying technologies, BIM, and energy performance evaluation.
Overall, the proposed methodology represents a meaningful step toward the automation of BIM-based workflows, contributing to more efficient, scalable, and intelligent processes that support sustainable design and the digital transformation of the built environment.
-
ROQUE, P. T.; DALLA VECCHIA, L. R. F.; BERTOLDI, N.; SILVA, A. C. S. B. da. Process of converting laser scans into BIM models for energy simulation purposes. Ambiente Construído, Porto Alegre, v. 26, e151650, jan./dez. 2026. ISSN 1678-8621 Associação Nacional de Tecnologia do Ambiente Construído. http://dx.doi.org/10.1590/s1678-86212026000100975
-
Declaration of Generative AI and AI-Assisted Technologies in the Writing Process
The authors used generative AI tools to assist in language revision and improvement of the manuscript’s clarity and readability. The authors carefully reviewed and edited the content and take full responsibility for the final version of the manuscript.
-
Financial Support
The first author of the study was financed by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).
Data Availability Statement
Data will be made available on reasonable request.
References
-
APPLE. Capturing depth using the LiDAR camera 2024. Available: https://developer.apple.com/documentation/avfoundation/additional_data_capture/capturing_depth_using_the_lidar_camera Access: 08 out. 2024.
» https://developer.apple.com/documentation/avfoundation/additional_data_capture/capturing_depth_using_the_lidar_camera -
APPLE. Introducing RoomPlan Apple Worldwide Developers Conference, 2022. Video. Available: https://developer.apple.com/videos/play/wwdc2022/10077/ Access: Jun. 1, 2025.
» https://developer.apple.com/videos/play/wwdc2022/10077/ -
AUTODESK. 3D laser scanning: unleashing the power of precision. 2024. Available at: https://www.autodesk.com/solutions/3d-laser-scanning Access: 12 set. 2024.
» https://www.autodesk.com/solutions/3d-laser-scanning - BERTOLDI, N. D. Scan-to-BIM Sistematizado: proposta de processo utilizando sensor LiDAR em dispositivos móveis para simulação de eficiência energética. Pelotas, 2025. Master’s Dissertation – Graduate Program in Architecture and Urbanism, Universidade Federal de Pelotas, Pelotas, 2025.
- BORKOWSKI, A. S.; KUBRAT, A. Integration of laser scanning, digital photogrammetry and BIM technology: a review and case studies. Eng, v. 5, n. 4, p. 2395-2409, 2024.
- BUNGE, M. Epistemologia São Paulo: T.A. Queiroz Editora, 1980.
- CUPERSCHMID, A. R. M.; NEVES DE OLIVEIRA, G.; FRONER, Y.-A. Exploring the use of LiDAR in smartphones: documenting the frontispiece of Saint Francis of Assisi Church in Ouro Preto, Brazil. International Journal of Architectural Heritage, v. 19, n. 9, p. 1397–1414, 2024.
- DRESCH, A.; LACERDA, D. P.; ANTUNES, J. A. V. Design science research: a method for science and technology advancement. Porto Alegre: Springer, 2014.
- GLEŃ, P.; KRUPA, K. Comparative analysis of the inventory process using manual measurements and laser scanning. Budownictwo i Architektura, Lublin, v. 18, n. 2, p. 21-30, 2019.
-
HEVNER, A.; PARK, J.; MARCH, S. T. (ed.). Design science research in information systems. January 20, 2004 (created in 2004 and updated until 2015 by Vaishnavi, V. and Kuechler, W.); last updated (by Vaishnavi, V. and Petter, S.), June 30, 2019. Available: https://doi.org/10.1007/978-1-4419-5653-8_2 Access: Mar. 26 2026.
» https://doi.org/10.1007/978-1-4419-5653-8_2 - INTERNATIONAL ORGANIZATION FOR STANDARDIZATION. ISO 16739-1: Industry Foundation Classes (IFC) for data sharing in the construction and facility management industries: part 1: data schema. Geneva, 2024.
- LACERDA, D. P. et al. Design science research: método de pesquisa para a engenharia de produção. Gestão & Produção, v. 20, n. 4, p. 741–761, 2013.
- MACHER, H.; LANDES, T.; GRUSSENMEYER, P. From point clouds to building information models: 3D semi-automatic reconstruction of indoors of existing buildings. Applied Sciences, v. 7, n. 10, p. 1030, 2017.
- PIMENTEL, B. P. et al. Comparação dos métodos simplificado e de simulação propostos no novo regulamento brasileiro de etiquetagem de edificações públicas. Ambiente Construído, Porto Alegre, v. 21, n. 4, p. 179–200, out./dez. 2021.
- RODRIGUES, C. M. A. C. A tecnologia BIM como ferramenta de simulação para etiquetagem de edifícios conforme a INI-C Pelotas, 2023. Master’s Dissertation – Graduate Program in Architecture and Urbanism, Universidade Federal de Pelotas, Pelotas, 2023.
- SADEGHINEKO, F.; LAWANI, K.; TONG, M. Practicalities of incorporating 3D laser scanning with BIM in live construction projects: a case study. Buildings, v. 14, n. 6, p. 1651, 2024.
- VENABLE, J. R. The role of theory and theorising in design science research. DESRIST, v. 24-25, p. 1-18, 2006.
- WANG, Q.; GUO, J.; KIM, M. K. An application oriented scan-to-BIM framework. Remote Sensing, v. 11, n. 3, p. 365, 2019.
- ZENG, R. et al Integrating as-built BIM model from point cloud data in construction projects. Engineering, Construction and Architectural Management, v. 31, 2023.
Edited by
-
Editor in-chief:
Enedir Ghisi
-
Guest editor:
Fernando Sá Cavalcanti
