Skip to Main Content
Traffic surveillance has increasingly become an important research topic in Intelligent Transportation Systems (ITS). Recently, a number of traffic monitoring systems using Closed Circuit Televisions (CCTVs) have been implemented in many urban areas all over the world in order to monitor, record image sequences and report traffic information. Images are captured from cameras and are sent to other systems for processing or displaying at the traffic management center or websites. However, data transmission problems and/or camera malfunction can corrupt the quality of video images. In this paper, we propose a simple but efficient method to detect corrupted images received from traffic surveillance cameras using histogram analysis and neural network classifiers. The experimental results show that our approach can accurately differentiate between normal and corrupted images.