| Project Title | Automated FHWA Vehicle Classification Using Combined Semantic and Geometric Features Collected from |
| University | West Virginia University |
| Principal Investigator(s) | Fei Dai |
| Funding Source(s) and Amounts Provided (by each agency or organization) | Federal Funds, $127,610 Match, $127,610 |
| Total Project Cost | $255,220 |
| Start and End Dates | 05/01/2023 - 07/31/2025 |
| Brief Description of Research Project | The objective of this project is to develop an automated vehicle classification method following the FHWA 13-category classification schema through measurement conducted on traffic surveillance videos. Different from the existing methods, the proposed method will exploit the combined semantic and geometric features extracted from the traffic video frames for data processing and information acquisition. Supported through the CIAMTIS core project on “image-based vehicle height measurement for prevention of low clearance infrastructure collisions”, the PI has done extensive work on geometric measurement for vehicle height capture via deep earning, computer vision and single view geometry. This proposed project will be an expansion of the existing work that will focus on vehicle axle configuration identification and encoding that allows for computer to automatically recognize fine-grained truck classes using the classification rules defined by FHWA. |
