Traffic event detection from road surveillance vide os based on fuzzy logic

TLDR: A novel Fuzzy Logic based analysis framework and a video based traffic data extraction scheme to decide upon the right traffic conditions are presented and are seen to be robust enough to reject the noisy data coming from surveillance videos.

Abstract: This work is about the autonomous detection of a road traffic incident by exploiting road surveillance camera videos. Timely and autonomous detection of an incident is paramount for the reduction of traffic congestion so that countermeasures can be taken at the earliest. This paper presents a novel Fuzzy Logic based analysis framework and a video based traffic data extraction scheme to decide upon the right traffic conditions. The existing road traffic analysis approaches as reported in literature do not extract data from the road camera videos; rather they use already available data to validate their schemes. However, in the proposed approach a complete scheme is proposed which takes a raw road camera video and autonomously extracts the relevant data for the subsequent Fuzzy Logic based traffic analysis. To show the efficacy of the proposed scheme, unprocessed surveillance videos of both urban and motorway scenarios are used. The results indicate that the traffic flow and their statistics are adequately determined through the proper selection of membership functions and rule formulation. Owing to the use of fuzzy logic, our proposed framework is seen to be robust enough to reject the noisy data coming from surveillance videos.