As-Built Modelling and Assessment

Size: px

Start display at page:

Download "As-Built Modelling and Assessment"

Baldric Joseph
5 years ago
Views:

1 As-Built Modelling and Assessment Ioannis Brilakis, Ph.D. Laing O Rourke Lecturer in Construction Engineering Civil, Structural and Environmental Engineering

Restoring and Improving Urban Infrastructure Engineering Grand Challenge (NAE, 2008) Lack of viable methods to map/label existing infrastructure E.g. 2/3 of modelling effort spent manually converting raw data to geometric model (Jaselskis et al.

2 Restoring and Improving Urban Infrastructure Engineering Grand Challenge (NAE, 2008) Lack of viable methods to map/label existing infrastructure E.g. 2/3 of modelling effort spent manually converting raw data to geometric model (Jaselskis et al., 2008) Result: As-built models not produced for most new and retrofit construction Leads to rework and design changes (NAE, 2008) Cost up to 10% of installed costs (Reddington, 2005) Lead to reduced sustainability (significant material waste)

4 Outline CAREER Award # First step Basic Research Examples Infrastructure Materials Recognition Infrastructure Recognition Measure Object Size Detect Object Defects Count Track Roof Dimensions, Excavation Volume Bricks, Doors, Workers, Trucks Materials, Personnel, Vehicles, Equipment Cracks, Air pockets, Spalling, Potholes Application Areas As-Built Geometric Modelling As-Built Quantity Take-Offs Compare As-designed / As-built Productivity Measuring Sustainability Monitoring Project Monitoring Systems Congestion Control Systems Activity Sequence Analysis Structural Damage Assessment Maintenance Decision Making

5 Material Based Image Classification Clustering Shape Information Material Signature concrete insulation Signature Comparison Image {materials} {shapes} Image Repository KBase steel roofing steel

Multi-modal Image Retrieval Model Based System Object attributes query Materials = bricks, insulation Month = October Year = 2004 X_location = 145 m Y_location = 12 m

6 Multi-modal Image Retrieval Model Based System Object attributes query Materials = bricks, insulation Month = October Year = 2004 X_location = 145 m Y_location = 12 m Z_location = 5 m Construction Database Images Retrieved Attribute Comparison Image attributes Materials = bricks Day = 23 Month = October Year = 2004 Image Repository

7 7

8 Infrastructure Recognition CAREER Award # First step Basic Research Examples Infrastructure Materials Recognition Infrastructure Recognition Measure Object Size Detect Object Defects Count Track Roof Dimensions, Excavation Volume Bricks, Doors, Workers, Trucks Materials, Personnel, Vehicles, Equipment Cracks, Air pockets, Spalling, Potholes Application Areas As-Built Geometric Modelling As-Built Quantity Take-Offs Compare As-designed / As-built Productivity Measuring Sustainability Monitoring Project Monitoring Systems Congestion Control Systems Activity Sequence Analysis Structural Damage Assessment Maintenance Decision Making

9 From spatial/visual data to a 3D Model Point clouds / images have no knowledge of the elements they contain What they are made of Which points/pixels belong to which entities Cannot provide any other as-built information besides spatial measurements i.e. material info, health info, etc. Purpose of as-built modelling = Convert spatial/visual data to information rich, object-oriented model

Infrastructure Recognition CAREER Award #0948415 VPR Models 3D Model For each representative

Knowledge Image Analysis Model-based Recognition Identify its visual characteristics

11 Infrastructure Recognition CAREER Award # VPR Models 3D Model For each representative element: Recognize & Detect Images Remote Sensing Match in Point Cloud 3D Points Domain Knowledge Image Analysis Model-based Recognition Identify its visual characteristics Identify characteristics representation features Create element model Add to database of element models

12 Results Columns Detection CAREER Award # % precision (TP/TP+FP) 79.1% recall (TP/TP+TN) 320x240, 25 fps on a netbook

13 As-Built Geometry CAREER Award # First step Basic Research Examples Infrastructure Materials Recognition Infrastructure Recognition Measure Object Size Detect Object Defects Count Track Roof Dimensions, Excavation Volume Bricks, Doors, Workers, Trucks Materials, Personnel, Vehicles, Equipment Cracks, Air pockets, Spalling, Potholes Application Areas As-Built Geometric Modelling As-Built Quantity Take-Offs Compare As-designed / As-built Productivity Measuring Sustainability Monitoring Project Monitoring Systems Congestion Control Systems Activity Sequence Analysis Structural Damage Assessment Maintenance Decision Making

collection + post processing man-hours) Result: Small/medium sized on-going and completed

14 Real World Infrastructure to 3D Points Measuring object size needs 3D data Getting adequate/accurate 3D data is costly Device costs (i.e. laser scanner) + Labour costs (data collection + post processing man-hours) Result: Small/medium sized on-going and completed projects cannot afford it Let alone getting 3D data frequently, throughout the construction phase

15 Coordinating Efforts Marie Curie IRSES Program Consortium established in 2008 to formalize the as-built modeling process (video to BIM), coordinate efforts

16 Damage and Defects Assessment CAREER Award # First step Basic Research Examples Infrastructure Materials Recognition Infrastructure Recognition Measure Object Size Detect Object Defects Count Track Roof Dimensions, Excavation Volume Bricks, Doors, Workers, Trucks Materials, Personnel, Vehicles, Equipment Cracks, Air pockets, Spalling, Potholes Application Areas As-Built Geometric Modelling As-Built Quantity Take-Offs Compare As-designed / As-built Productivity Measuring Sustainability Monitoring Project Monitoring Systems Congestion Control Systems Activity Sequence Analysis Structural Damage Assessment Maintenance Decision Making

17 Post-Earthquake Structural Safety Assessment FEMA 306 (rapid) ATC-20 (rough) FEM (detailed)

Problem Statement & Objectives Problem Earthquakes create huge inspection demand within seconds, takes weeks to months to assess all structures In the mean time Objective ERs risk

18 Problem Statement & Objectives Problem Earthquakes create huge inspection demand within seconds, takes weeks to months to assess all structures In the mean time Objective ERs risk their lives in unsafe buildings (FEMA codes) People stay out waiting for assessment (ATC codes) Shift the research focus from accuracy to speed Get useful measurements automatically

Rapid Safety Assessment Grant #1000700 Co-PIs: R. DesRoches, L.

bearing capacity Video Frames Processed Scene Parts Vision Tracking

W* Δ - Max W / W* Average 3.29 2.21% 0.35% Std 2.70 2.90% 0.

Detection Column Damage Assessment h j l k i m Damage Properties

19 Rapid Safety Assessment Grant # Co-PIs: R. DesRoches, L. Lowes Get rapid measurements from video Translate them to remaining bearing capacity Video Frames Processed Scene Parts Vision Tracking New Scene Parts Warn if collapse is imminent Δ - Orientation Δ - L / W* Δ - Max W / W* Average % 0.35% Std % 0.49% Video camera Concrete Columns Damage Detection Concrete Columns Detection Column Damage Assessment h j l k i m Damage Properties Column Damage Index f a c b e d g PDA Fragility Curve Database Risk Collapse Probability Assessment

20 and there is much left to do CAREER Award # First step Basic Research Examples Infrastructure Materials Recognition Infrastructure Recognition Measure Object Size Detect Object Defects Count Track Roof Dimensions, Excavation Volume Bricks, Doors, Workers, Trucks Materials, Personnel, Vehicles, Equipment Cracks, Air pockets, Spalling, Potholes Application Areas As-Built Geometric Modelling As-Built Quantity Take-Offs Compare As-designed / As-built Productivity Measuring Sustainability Monitoring Project Monitoring Systems Congestion Control Systems Activity Sequence Analysis Structural Damage Assessment Maintenance Decision Making

21 Future Research Directions Automation solutions for improved productivity and sustainability in construction Building objects detection and inference methods Sustainable pavement assessment methods As-is bridge modeling (BrIM) / highway asset mapping Vehicle classification and tracking for congestion pricing, stop-and-go waves estimation, etc. Automation advances in structural safety inspection UAV navigation for post-disaster structural assessment

Thank you! Questions? This material is based upon work supported by the National Science Foundation under Grants No. 0625643, 0738417, 0800170, 0943112, 0948415, 1000700,1031329, and 1034845.

22 Thank you! Questions? This material is based upon work supported by the National Science Foundation under Grants No , , , , , , , and It is also supported by the European Commission under Grant FP7:IRSES: BIMAutoGen, the German Academic Exchange Service (DAAD), the Georgia Department of Transportation and Metalforming Inc. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the sponsors named above.

Address: 103 East Architecture Building Georgia Institute of Technology th Street Atlanta, GA Tel:

Zhenhua Zhu Ph.D. Candidate School of Civil and Environmental Engineering Address: 103 East Architecture Building 245 4 th Street Atlanta, GA 30332 Tel: 404-385-1276 Email: zhzhu@gatech.edu RESEARCH INTERESTS