Grants and Contracts Details
Abstract OHS, Assessing and Extracting Speed Limit Data through Machine Learning 10/01/2021 – 09/30/2022 PI: Ben Blandford KYTC currently has posted speed limit data for all state maintained roads and some local roads. This project would seek to improve the completeness and accuracy of the speed limit data by 1) Performing a quality check analysis of the current speed limit data, and 2) Identifying posted speed limits for roads where the data are currently missing in the system. To accomplish this, the project would develop a machine learning methodology for analyzing Google Streetview Imagery and systematically identifying speed limit signs along roads. Data obtained from the machine learning process would be mapped and compared to speed limit data currently in the KYTC HIS database. Discrepancies would be flagged and presented to KYTC for review and potential update. The KTC research team will adapt a machine learning methodology for processing Google StreetView images to learn to recognize speed limits signs located along the roadway. Once the methodology is established, the team will validate the machine learning process by manually checking a subset of the speed limit data results produced for accuracy. The team will access StreetView photos through the API and apply the methodology to the photos in a systematic manner across the state. For roads that KYTC already has speed limit data for, results will compared and any discrepancies will be flagged for further investigation. For roads that KYTC does not currently have speed limit data for, the newly produced data will be spatially referenced in LRS and delivered to KYTC. 1. To verify the accuracy of all speed limit data currently available for Kentucky roadways by September 30, 2022. 2. To increase the completeness of the speed limit data by applying the machine learning methodology to all Kentucky roads available through Google StreetView by September 30, 2022.
|Effective start/end date||10/1/21 → 9/30/22|
- KY Office of Highway Safety: $69,721.00
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.