A Framework for Model-Driven Acquisition and Analytics of Visual Data Using UAVs for Automated Construction Progress Monitoring

Real-Time 3D Reconstruction on Construction Site Using Visual SLAM and UAV

“…With onboard cameras, UAV can take aerial images during flight and produces 3D point clouds through image-based postprocessing techniques (Liu et al 2014). Recent applications showed the potentials of using UAV to monitor construction progress (Lin et al 2015), survey civil infrastructures (Chan et al 2015), and document historical site (Themistocleous et al 2016). However, due to the heavy computational workload of image processing and computer vision algorithms, the model generation process is still slow (Nex and Remondino 2014), which limits its implementation on time critical applications.…”

Visual monitoring of civil infrastructure systems via camera-equipped Unmanned Aerial Vehicles (UAVs): a review of related works

“…Their system, similar to (Karsch 2999) overlays color-coded 3D construction progress information on the physical world and adopts filtering methods to avoid information clutter and issues associated with displaying detailed augmented information. Other recent examples by (Han et al 2015;Lin, Han, Fukuchi et al 2015a) introduce web-based tools with scalable system architectures for visualizing and manipulating large scale 4D point cloud data, large collections of images, 4D BIM, and other project information. These methods account for level of details in data representation and dynamically consider the limited computational power and connection bandwidth for visualizing data on commodity smartphones.…”

“…It also requires techniques for as-built and as-damaged information modeling; nevertheless, techniques that address these issues are not well studied. Ongoing efforts such as the Flying Superintendents project at the University of Illinois (Lin, Han & Golparvar-Fard 2015), and the ARIA project at Carnegie Melon University (ARIA (Team 2015)) are geared towards providing specific frameworks together with methods for construction monitoring and civil infrastructure condition assessment to address some of these gaps-in-knowledge.…”

“…1 A vision for the next generation construction site where camera-equipped UAVs autonomously monitor work-in-progress for improving safety, quality, and productivity information; and (3) identification of the most informative views for data collection (e.g., canonical view, top down view, etc. for appearance-based recognition of work-inprogress as in (Han et al 2015)). Table 1 categorizes prior works based on their level of autonomy and with respect to their specific applications.…”

“…While these methods generally work better on ordered and sequential images, yet without a rigid control on the location and viewpoint of the captured images with respect to onsite construction elements, the visual content of the images may not be most suitable for certain applications. For example, for appearance-based construction progress monitoring, leveraging material recognition techniques such as Han et al (Han et al 2015) benefits from images that are taken in a canonical view to identify the most updated status of construction for onsite elements. This requires research that identifies the most informative views, and then techniques that can leverage such information as a priori information for data collection.…”