Browsing by Author "Sornsakul, Chavisa"

Now showing 1 - 1 of 1
  • Thumbnail Image
    ItemOpen Access
    Improving Sidewalk Maintenance through Smartphone Citizen Reporting of Pothole Damage
    (2024-10-18) Sornsakul, Chavisa; Lichti, Derek; O'Keefe, Kyle; Radovanovic, Robert; Detchev, Ivan
    Pedestrian walkways potholes pose significant risks to pedestrian safety, causing accidents and disruptions. This thesis explores using Structure-from-Motion (SfM) 3D reconstruction with smartphone GNSS measurements to improve citizen pothole reporting, aiming to enhance local engagement, infrastructure management, and public safety. The methodology uses SfM camera poses recovered from citizen pothole image reports and relative positioning from smartphone GNSS measurements to resolve the scale ambiguity of SfM pothole models. The relationship between images and a reconstructed point cloud from the SfM technique estimates pothole quantity, perimeter, and maximum depth. Key contributions include developing an integrated system for accurate pothole dimension measurement, with a scale error of approximately 12% from reference measurements, and establishing data acquisition guidelines for assessing the reliability of citizen-reported data. This method results in pedestrian walkway pothole perimeters differing by approximately 35 cm and depths by less than 1 cm. A guide for citizens capturing pothole images is proposed.