A majority of traffic counting in the U.S is performed using portable traffic counters employing a road tube system that is unsafe, difficult to operate, and unable to meet the data quality standards desired by transportation agencies. This project developed a portable traffic monitoring device for counting and Federal Highway Administration’s 13 vehicle classification, the standard vehicle classification system in the U..S. The device is a video-based system that uses a combination of artificial intelligence and image processing to collect data on a single lane. The performance evaluation results indicate consistent counting and classification with more than 98% accuracy. Return on investment analysis is presented, taking into account material costs, productivity gain and collection rejection rates. The findings indicate that the developed device has the potential for wide use in the industry. There are over 100 units currently in use by transportation agencies to collect traffic monitoring data.

The report is available.