Traffic characteristics study through processing of video image for expressway in india

Abstract

India has the third-largest road network in the world. Expressways are the highest class of roads in the Indian Road Network. India’s progress in the road sector measure 600 km of expressways approximately and is planning for achieving more than 15000km of expressway by 2021. Expressways are a controlled-access highway designed for fast traffic, with controlled entrance and exit. Expressways are vastly different from other roads of the country as vehicles such as bicycles, two-wheelers, three-wheelers and bullock carts are not allowed on these roads and additionally, there is no strict lane discipline. A number of research papers are available on studies on traffic stream characteristics for various roadway and traffic conditions. However, very limited studies are carried out on expressways in India Video graphic data collection is widely used in traffic engineering, as video recordings can act as a more detailed, complete, accurate and reliable observational technique. To figure out the exact relationship between the traffic parameters, lots of research has been done over the past. Many attempts have been made earlier for data retrieval from videos. But, there have been very few attempts to discuss the methodology for data extraction and analysis from video for expressway data in India. This paper discusses about video graphic data collection, extraction and analysis of traffic stream characteristics for a duration of 8 hours by taking Ahmedabad -Vadodara expressway (four lane divided carriage way) as a case study. Data has been collected using video graphic survey by fixing high resolution cameras on road over bridge (ROB) such that traffic flow faces the camera. Data on traffic volume, vehicle composition, speed of different vehicle categories, lane utilization etc. are manually extracted from the recorded video after converting it into frames. Data extraction and analysis has been done with the help of the softwares such as Ulead Video Studio, Irfan View and SPSS, MS Excel. The analysis on degree of lane discipline showed that, 97% of vehicles are following lanes. The results also showed that median side lane had high speed traffic as compared to other lane, which may be due to very high car composition (90%).