Fog- and weather-related visibility reduction is a common cause of multiple-vehicle crashes. Large differential speeds and a tendency of vehicle operators to drive faster than is safe can lead to terrible crashes. Fog can usually be seen on traffic cameras, which are becoming more prevalent on Virginia highways as well as on highways in other states. This project studied the applicability of one approach to using machine vision to measure fog in a realistic environment simulated on the Virginia Smart Road. With the assistance of Dr. Eric Dumont, a leading visibility research from IFSTTAR in France, a machine vision algorithm was applied to video stills captured from a common traffic camera installed on the Smart Road. Machine vision algorithms were used to determine the average loss in visual detail in the scene viewed by the camera and this was used to generate an empirical model relating Meteorological Optical Range (MOR) and the camera images. The model was used to evaluate data captured on days close in time and days over the following year. Finally, the research investigated the approach’s sensitivity to preset positioning errors in the camera. The research shows that the approach has promise. However, further research and development are needed before the approach is ready for deployment.