Computer Vision Measurements and Analysis to Support Naturalistic Driving Studies for Road Safety

The project is divided into five phases. In Phase 1, Oak Ridge National Laboratory (ORNL) will design a calibration procedure for instrumented vehicles, then acquire and process calibration data from the second Strategic Highway Research Program (SHRP2) 24-car study instruments. Phase 2 has two subphases consisting of the investigation of infield camera calibration methods (subphase 1) and the development of baseline Automated Identity Masking (AIM) and Automated Feature Extraction (AFE) modules (subphase 2). In Phase 3, ORNL will test the baseline AIM and AFE modules and create an AIM-effectiveness metric based on biometric and feature extraction similarity. In Phase 4, ORNL will conduct the initial evaluation of PT code with PT assistance. A larger data set from the full SHRP2 Naturalistic Driving Study (SNDS) will be created for blind evaluation of algorithms. In conjunction with this blind test set, ORNL will perform additional calibrations on SNDS vehicles as available and accessible. The Oak Ridge National Laboratory will also create computational load projections for PT algorithms of interest and begin design of a data-sharing architecture. Finally, in Phase 5, ORNL will perform final comparisons and evaluations of PT algorithms. The Oak Ridge National Laboratory will develop a plan for future SNDS feature extraction grand challenges and data analytics challenges in conjunction with a data-sharing architecture.