Visual and Acoustic Data Analysis for the Bridge Deck Inspection Robotic System

Abstract

Abstract - Bridge deck inspection is essential task to monitor thehealth of the bridges. This paper reports the data collectionand analysis for bridge decks based on our novel roboticsystem which can autonomously and accurately navigateon the bridge. The developed robotic system can lessen thecost and time of the bridge deck data collection and risksof human inspections. The advanced software is developedto allow the robot to collect visual images and conductnondestructive evaluation (NDE) measurements. The imagestitching algorithm to build a whole bridge image fromindividual images is presented in detail. The impact-echo(IE) and ultrasonic surface waves (USW) data collected bythe robot are analyzed to generate the delamination andconcrete elastic modulus maps of the deck. Keywords - Mobile robotic systems, Bridge deck inspection, ImageStitching, Nondestructive evaluation. I. Introduction The condition of bridges is critical for the safetyof the traveling public and economic vitality of thecountry. There are many bridges through the U.S. that arestructurally deficient or functionally obsolete. Conditionmonitoring and timely implementation of maintenanceand rehabilitation procedures are needed to reduce futurecosts associated with bridge management. Applicationof nondestructive evaluation (NDE) technologies is oneof the effective ways to monitor and predict bridgedeterioration. A number of NDE technologies are cur-rently used in bridge deck evaluation, including impact-echo (IE), ground penetrating radar (GPR), electricalresistivity (ER), ultrasonic surface waves (USW) testing,visual inspection, etc. [5], [22]. For a comprehensive andaccurate condition assessment, data fusion of simultane-ous multiple NDE techniques and sensory measurementsis desirable. Automated multi-sensor NDE techniqueshave been proposed to meet the increasing demandsfor highly-efficient, cost-effective and safety-guaranteedinspection and evaluation [7].Automated technologies have gained much attentionfor bridge inspection, maintenance, and rehabilitation.Mobile robotic inspection and maintenance systems aredeveloped for vision based crack detection and main-tenance of highways and tunnels [18], [19], [23]. Arobotic system for underwater inspection of bridge piersis reported in [3]. An adaptive control algorithm for abridge-climbing robot is developed [15]. Additionally,robotic systems for steel structured bridges are developed[2], [16], [21]. In one case, a mobile manipulator is usedfor bridge crack inspection [20]. A bridge inspectionsystem that includes a specially designed car with arobotic mechanism and a control system for automaticcrack detection is reported in [11], [12], [17]. Similarsystems are reported in [13]–[15] for vision-based auto-matic crack detection and mapping and [24] to detectcracks on the bridge deck and tunnel. Edge/crack detec-tion algorithms such as Sobel and Laplacian operatorsare used.Difference to all of the above mentioned works, ourpaper focus on the bridge deck data analysis which iscollected by our novel robotic system integrated withadvanced NDE technologies. The developed data analy-sis algorithms allows the robot to build the entire bridgedeck image and the global mapping of delamination andelastic modulus of the bridge decks. These advanceddata analysis algorithms take into account the advantagesof the accurate robotic localization and navigation toprovide the high-efficient assessments of the bridgedecks.The paper is organized as follows. In the nextsection, we describe the robotic data collection sys-tem and coordinate transformation. In Section III wepresent the image stitching algorithm and bridge deckviewer/monitoring software. In Section IV, we presentThe 31st International Symposium on Automation and Robotics in Construction and Mining (ISARC 2014)