Discrimination of highway snow condition with video monitor for safe driving environment

Detection of road condition is important for safe driving and secure transportation. In northern China, snow season could last for half of the year. Knowledge of highway condition is used as a trigger for snow removal or road closure. Conventional discrimination for road conditions using optical or ultrasonic sensors. However, these sensors could only provide point information, which usually could not reflect the spacious condition. To deal with this problem, we proposed a snow detection scheme with video monitor. Road texture feature is extracted from the selected areas with co-occurrence matrix. Then we compare several discrimination classifiers for “Heavy Snow Cover”, “Mild Cover” and “Dry” detection. Experimental results show that the best discrimination accuracy could achieve above 93%.