Virtual and remote inspection methods Rasmus Lindeneg Johansen, RLJ@cowi.dk 1
How do we inspect "today"? 2
Inspection "today" Manual inspections From the ground From lift/platform By rope Data was collected On paper Or digitally (Database or FieldNote-app) Challenges: Require traffic restrictions, High initial costs, Papers gets wet, fly away etc. Very dependent on weather Slow/requires long time in the field Difficult to document 3
A "new" approach 4
Virtual Inspection Use COWI Virtual Inspection (CVI) to perform "virtual inspections" for initial screening of the structure Continued need for manual inspection in the field of selected areas (typically less than 10%) FACTS about CVI: CVI is a web-based solution Digital images are collected by UAVs, handheld cameras or panoramic cameras Observations are annotated directly on the 3D-model 5
Virtual Inspection FACTS about CVI: User-defined inspection report format Able to add historical data including images, reports and pdf drawings. Development of automatic detection methods using AI / Machine Learning is ongoing. Systematic inspection using a sweep plan. Drag / drop for import of own "georeferenced" data 6
3D model of a bridge anchorblock based on images collected by drones Image from 360 degree camera 7
Inspection is done in high resolution images. Markings are added to all images and 3D model 8 Findings are added to a database
9 Export example. Format pdf, word and excel
10 Images and drawings can be added (= data management system)
Virtual and remote inspection methods for bridges CVI is scalable.. 11
Example: "Virtual inspection" of wells 12
Example: "Virtual inspection" of building 13
Future improvement and perspectives 14
15 2 pilots 3 UAV's (one in reserve) 6 hours flight 9.000 images
How is it made? Images are automatically aligned through common tie point calculation all images are referenced to each other Automatic point cloud calculation and manual clean-up high quality point cloud, low noise level Manual 3D model creation and automatic texturing Low polygon count, looking good 16
UAV - Platform Phantom 4 Pro First Anchor block completed using standard DJI Phantom 4 Pro Unable to look up and low GPS accuracy (1-2 meters) Matrice 210 RTK Option to mount camera on top and look up underneath the bridge better GPS positional accuracy through RTK (15-20 cm) 17
Automated UAV-Flight path With the processed overall 3D model as planning basis the exposure points for the systematic and detailed image acquisition is calculated and then the automated and systematic image acquisition of the detailed images is ready 18
Future perspective BIM integration Add the laser scanned interior to the 3D model Terrestrial scans Better integration of point clouds Deep Learning automated image analysis Link to existing information create and display information from existing databases 19
How can CVI create value for you? 20
Main points We create a platform for easy access to information and documentation COWI's own on-line web platform, it's here, it's working and we are ready It's scalable and adjustable Fairly short turn around, from acquisition to published data in 1-2 weeks Pre-calculated automatic flight paths can be re-used and provide images from same location year after year for documentation of changes over time Combine multiple data sources 3D model, BIM, aerial imagery, point cloud, vector maps, pdf's, etc. 21
How can CVI create value for you? CVI can be used during planned inspection campaign for management and inspection, as well as for handling of data for the "unplanned" visits (drag and drop solution) Historical data can be integrated = one platform for viewing of data Use data for BIM modelling and for design purposes The Engineer can use the data for planning of site trip Data can be shared with Consultants, Architects and Contractor involved in specific task 22
23 RLJ@cowi.dk