BIMProve demonstrators covering stationary, ground- and UAV-based data acquisition systems; augmented reality (AR) visualisation of BIM model; application of multi-user virtual reality (VR); computer vision model for detection of safety structures in the construction site; mapping of point clouds to IFC elements; and BIMprove web front end
Augmented Reality (AR) visualisation of BIM Model
Description The demo shows BIM@Construction AR tool demonstrated with Android table and Microsoft HoloLens 2 devices.
Outcome The following functionalities have been demonstrated: BIM Model visualisation in AR in 1st and 3rd person point of view.
Multi-User Virtual Reality
Description The demo shows the different functionalities already implemented in the MUVR-Viewer.
Outcome The following functionalities have been implemented and tested successfully for both PC- and VR-users: a) automatically loading 3D-geometry (including point cloud-data) and deriving a user-menu from it; b) connecting to a virtual room and joining a multi-user-session; c) (de-)selecting 3D-models to show/hide them; d) 1st rudimentary version of issue management; and e) means of remote communication.
Ground Robot BIMprove system integration and operation
Description The demo shows how the localization and navigation on ground robots works, by illustrating the integration between Bimsync and Robotnik's Backend, in which one the robot automatically gets both 2D and 3D model and use them to locate and navigate on site.
Outcome The use of 2D-3D BIM models for ground robot navigation and localization.
Leica BLK 360 Point-clouds scans
Description The demo shows the Leica BLK360 laser scanning the scene by obtaining the point cloud, heat map and high resolution images.
Outcome Through the SW Cyclone the laser scanning information is obtained and allows to export it to e57 for its interpretation.
Data-capturing drones
Description The demos shows the indoor helicopter flight with on-board sensors, gimbal camera for data capture, and the process of scanning at fixed distance and angle to wall.
Outcome A stable flight, flexible system, and extendable with other sensors/needs.
Detection of safety structures and untidy places in images
Description The demo shows the computer vision based analyser finds safety barriers, safety nets and untidy places from the construction site images. These are pointed out with rectangles in the images. Health and safety inspections are required at construction sites to prevent accidents including falling from heights and fire hazards. In BIMprove, safety inspections will be supported by computer vision.
Outcome Computer vision model for detecting safety barriers, safety nets, and untidy places in received images. Web UI for viewing the results.
Mapping point clouds to IFC-elements
Description The demos shows the classification of which points in a cloud that belongs to a certain IFC-element.
Outcome Two point clouds, one consisting of all points mapped to an ifc-element and the second is all points to mapped to an ifc-element.
Revision change visualisation
Description The demo shows BIMSync Backend register revisions update of IFC-file, which produces a difference file between original and new revision.
Outcome Colored IFC-file illustrating the difference and a BCF issue list.
Schedule import and view
Description The demo shows the import schedule from Microsoft Project native format into a database, and this data in a Gantt diagram on a web page.
Outcome Import and viewing of schedule works with MPP files.
BIM@Vehicle: First Version of the UI
Description The demo shows BIM@Vehicle tools UI concept as normal navigator like 2D UI. User could enhance it by link it to AR tool, in this case Microsoft HoloLens 2 device.
Outcome Tools functionalities has been demonstrated in XR lab environment due to COVID-19 restrictions. User could visualized BIM Model and path to Point of Interest in 1st and 3rd person view. 1st person view has been done via AR with Hololens 2 and 3rd person view is exploiting Android tablet.
Ground-robot Data Capturing Request
Description The demo showcases a set of functionalities developed for the ground robotic vehicle that can be triggered in order to autonomously record a dataset.
Outcome • Data capturing autonomous mobile robot: The rover can navigate autonomously within a known map. • Extendable mission planning tool: The robot HMI is able to schedule the mission and command the needed actions in different points of interest without further user involvement. • Robot- Sensor payload integration: The robot can trigger and offer data captured by mounted sensors, as well as provide different options from such sensors without the need of additional interfaces.
Multi-User-Virtual-Reality (BIMprove XR-Viewer)
Description The demo showcases the XR Viewer being used by two VR-users and one desktop-user. The users in this demo are all physically in the same room; however, the users could use the system from different locations at the same time.
Outcome • Visual comparison of plan vs. scan: The possibility to view and inspect a direct overlay of scanned point clouds of the actual construction with as-designed BIM-models in a 1:1-scale. • Communication: To be able to communicate via the multi-user-/conferencing-function. • Reduce Travel: Not having to travel to the construction site, because a user can view and inspect the construction site in an almost real way.
IR Data Capture with Drone
Description This demo shows the capture of infrared images with a drone indoors. Here, a key challenge is to manage a stable flight inside a building without GPS data. The starting point is a 2D floor plan of a room where the recordings are to take place.
Outcome • Easy to redo the same mission many times, always with same path, same quality • Use of the acquired images for safety issues and documentation.
Spatial Data Capturing request for Drone
Description This demo shows the complete path from planning a drone mission, reviewing the mission data, the actual flight mission of data processing and finally the presentation of the results. The focus is on a mission in the outdoor area.
Outcome • Easy to redo the same mission many times, always with same path, same quality • Versatile use of the acquired images data: Photogrammetry, detecting safety issues, documentation of building progress.
Detecting safety nets
Description Risk Object Visual Analysis System, ROVAS, analyses images captured from the risk zones for detecting the existence and placement of the required safety structures.
Outcome • ROVAS that is able to analyse safety nets, barriers, and untidy places of images taken at construction sites • Result of the analysis is an image where the detected objects are marked with labelled bounding boxes and confidence level. • Dataset of 3500 labelled images has been used for training and validation • Mean average precision achieved for detecting safety nets in 960x960 images is 0.946.
