Leveraging Autonomous Vehicles In Transportation Asset Management

Traffic sign inventory generation is important for transportation asset management because accurate sign records support roadway safety, maintenance planning, and asset-level decision making. However, current field inventory, image-log, street-view, mobile mapping, and computer vision methods may be limited when agencies need to update sign records across roadway networks under practical data collection and processing requirements. This limitation makes it difficult to convert ordinary roadway observations into map-based traffic sign inventory records in a transportation asset management setting. To address this problem, this thesis proposes an integrated video-plus-GPS workflow for traffic sign inventory generation. The proposed workflow connects traffic sign detection with sign-region segmentation and combines monocular depth evidence with GPS-and-bearing coordinate projection so that duplicate-filtered traffic sign inventory records can be obtained. A case study using annotated road sign imagery and vehicle-mounted roadway video with GPS data is used to demonstrate and evaluate the proposed workflow. The results indicate that the detection stage can identify traffic signs across varied roadway scenes and that the combined segmentation, depth, coordinate prediction, and duplicate grouping stages can convert image observations into map-based inventory records. These findings suggest that the proposed workflow can support traffic sign inventory updating and maintenance planning for transportation agencies.

Name thesis:degree_name

Master of Science

Discipline thesis:degree_discipline

Engineering Data Science

Grantor

University of Houston

Year dc:date.issued

2026

Author dc:creator

• Veeramachaneni, Mahitha 2003-

Advisor dc:contributor.advisor

• Gao, Lu

Committee members dc:contributor.committeemember

• Kulkarni, Yashashree
• Mo, Yi-Lung

• Transportation asset management
• Traffic sign detection
• Monocular depth estimation
• Segment Anything Model
• RT-DETR
• Raffic sign inventory
• Vehicle-mounted video
• GPS-based geolocation

Language dc:language.iso

English

Handle dc:identifier.uri

https://hdl.handle.net/10657/21573

OAI identifier oai:identifier

oai:uh-ir.tdl.org:10657/21573