Image-Based Localization and Content Authoring in Structure-from-Motion Point Cloud Models for Real-Time Field Reporting Applications

Bae, Hyojoon; Golparvar-Fard, Mani; White, Jules

doi:10.1061/(asce)cp.1943-5487.0000392

Cited by 33 publications

(20 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Current methods such as (Garcia-Lopez and Fischer 2014) require significant upfront efforts to assemble the Work Breakdown Structure for work tracking. The documentation of what is DONE (Mossman 2013) and the knowledge of what SHOULD and CAN be done also relies on traveling between the site and trailers to access paper-based documents or, at best, searching on smartphones which requires 3D as-planned views to be manually generated for each task (Kamat et al 2010, Bae et al 2014. Sustaining these efforts for feedback and learning is key to improvement but is difficult to achieve.…”

Section: Related Workmentioning

confidence: 97%

Web-Based 4D Visual Production Models for Decentralized Work Tracking and Information Communication on Construction Sites

Lin

Golparvar-Fard²

2016

Construction Research Congress 2016

Self Cite

View full text Add to dashboard Cite

Tracking work-in-progress on construction sites requires establishing two-way flow of information between project managers and field personnel so that changes and performance issues with respect to task assignments, scope, status, and completion can be communicated effectively. Recent research has focused on minimizing decision bottlenecks and lowering coordination costs caused by inefficiencies in information sharing between project participants via leveraging 4D BIM and mobile computing platforms. Nevertheless, effective communication of who is doing what work in what location and comparing actual and planned progress in a timely fashion still requires preparing/updating 4D BIM with details that can reflect daily tasks, breaking the model into various locations, and then capturing task information onsite and manually updating them to 4D BIM. To address these inefficiencies, this paper presents a web-based work tracking system that builds on 4D visual production models, generated via superimposing image-based 3D point clouds with 4D BIM. These models allow project participants to directly commit to work tasks without the need for addressing difference in the level of detail between 4D BIM and actual work in progress, communicate performance deviations based on commitments from coordination sessions, measure readiness of the upcoming tasks based on inter-dependencies of current schedule tasks and their constraints, and highlight locations that are at-risk of potential performance problems. The benefits of the system in providing enhance awareness of ongoing work and better flow of information are discussed based on experiments conducted on a large-scale construction project. INTRODUCTIONThe increasing complexity of construction projects has made it extremely difficult to accurately predict key construction performance indicators, such as the project schedule and cost (Ham et al. 2016). It is no surprise that many owners are continuously looking for project delivery methods that provide more predictable project outcomes. Among the most important advancements in construction delivery methods that can bring accountability and predictability to project performance are lean methodologies and building information modeling (BIM). Hence, several construction companies have devised workflows and deliverables that integrate BIM and lean to help teams be more accountable as planners and builders. Nevertheless,

show abstract

Section: Related Workmentioning

confidence: 97%

Web-Based 4D Visual Production Models for Decentralized Work Tracking and Information Communication on Construction Sites

Lin

Golparvar-Fard²

2016

Construction Research Congress 2016

Self Cite

View full text Add to dashboard Cite

show abstract

“…Additionally, application of fiducial markers has been suggested by several researchers [24] but the systems are infrastructure-dependent and require the markers to be attached to various surfaces on construction sites. This challenges their applications for large-scale implementations [5]. Recent efforts presented a new vision-based system that allows tracking of the user's location and orientation by comparing images from the user's mobile device to a 3D point cloud model generated from a set of pre-collected site photographs [6].…”

Section: Computer Vision Technologiesmentioning

confidence: 99%

Infrastructureless approach for ubiquitous user location tracking in construction environments

Khoury

Chdid

Oueis

et al. 2015

Automation in Construction

View full text Add to dashboard Cite

“…For outdoor data sets we studied in [1][2][3][4][5], the value of N is typically in the range between 30,000 and 200,000 while the value of M is 10,000-20,000. Instead of time-consuming sequential matching, our new approach creates a single indexed k-d tree and uses a matching heuristic algorithm to find the target model for blind localization requests.…”

Section: Double-stage Matching Algorithm With a Single Indexed K-d Treementioning

confidence: 99%

“…After extracting the 3D point cloud of the subjects through a Structure-from-Motion (SfM) algorithm that estimates the 3D position of the visual features through image feature extraction, pair-wise matching, initial triangulation, and the Bundle Adjustment [25] optimization process, a 3D point cloud model can be used as a prior knowledge to compute 2D-to-3D correspondences for precisely localizing mobile camera imagery [26][27][28][29]. Using a 3D point cloud for user localization, i.e., model-based localization, permits mobile augmented reality systems to accurately estimate the 3D position and 3D orientation of the new photograph purely based on the image [1][2][3][4][5], and therefore, it does not have any hardware constraints on mobile devices, such as stereo cameras, GPS sensors, or motion tracking sensors. Furthermore, recent advances in SfM [30][31][32] enable the easy creation of large scale 3D point clouds from an unordered set of images and extend model-based localization methods to large scenes such as street-level or city-level scale.…”

Section: Related Workmentioning

confidence: 99%

“…The proposed mobile augmented reality approach, called as Hybrid 4-Dimensional Augmented Reality (HD 4 AR), that was first developed in our prior work [1][2][3][4][5], provides reliable identification of the location and orientation of the user based on photographs taken by existing and already available commodity smartphones. Videos of the commercial implementation of the technology by Cloudpoint Inc., are available on YouTube: https://www.youtube.com/user/PARworks.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Fast and scalable structure-from-motion based localization for high-precision mobile augmented reality systems

et al. 2016

Self Cite

View full text Add to dashboard Cite

A key problem in mobile computing is providing people access to cyber-information associated with their surrounding physical objects. Mobile augmented reality is one of the emerging techniques that addresses this problem by allowing users to see the cyber-information associated with real-world physical objects by overlaying that cyber-information on the physical objects' imagery. This paper presents a new vision-based context-aware approach for mobile augmented reality that allows users to query and access semantically-rich 3D cyber-information related to real-world physical objects and see it precisely overlaid on top of imagery of the associated physical objects. The approach does not require any RF-based location tracking modules, external hardware attachments on the mobile devices, and/or optical/fiducial markers for localizing a user's position. Rather, the user's 3D location and orientation are automatically and purely derived by comparing images from the user's mobile device to a 3D point cloud model generated from a set of pre-collected photographs. Our approach supports content authoring where collaboration on editing the content stored in the 3D cloud is possible and content added by one user can be immediately accessible by others. In addition, a key challenge of scalability for mobile augmented reality is addressed in this paper. In general, mobile augmented reality is required to work regardless of users' location and environment, in terms of physical scale, such as size of objects, and in terms of cyber-information scale, such as total number of cyber-information entities associated with physical objects. However, many existing approaches for mobile augmented reality have mainly tested their approaches on limited real-world use-cases and have challenges in scaling their approaches. By designing a multi-model based direct 2D-to-3D matching algorithms for localization, as well as applying a caching scheme, the proposed research consistently supports near real-time localization and information association regardless of users' location, size of physical objects, and number of cyber-physical information items. Empirical results presented in the paper show that the approach can provide millimeter-level augmented reality across several hundred or thousand objects without the need for additional non-imagery sensor inputs.

show abstract

Image-Based Localization and Content Authoring in Structure-from-Motion Point Cloud Models for Real-Time Field Reporting Applications

Cited by 33 publications

References 36 publications

Web-Based 4D Visual Production Models for Decentralized Work Tracking and Information Communication on Construction Sites

Web-Based 4D Visual Production Models for Decentralized Work Tracking and Information Communication on Construction Sites

Infrastructureless approach for ubiquitous user location tracking in construction environments

Fast and scalable structure-from-motion based localization for high-precision mobile augmented reality systems

Contact Info

Product

Resources

About