This report presents the research methodology, experiments, results, analysis, and future research directions for the automated collection of AusRAP road attributes using Digital Video Recorded (DVR) and pattern recognition techniques. The deep learning based 2-D segmentation and classification methods using DVR data for identifying AusRAP attributes are presented. The methods were used to develop new models and functions to detect AusRAP attributes. The models were evaluated on test data and incorporated into the development of software for automatically identifying AuSRAP attributes. The results and instructions on how to use the developed software are presented in this report.
Download ReportDownload Report
