G. HARISH BABU et al., International Journal of Emerging Trends in Engineering Research, 9(3), March 2021, 286 – 293 286 ABSTRACT The problem of traffic congestion has increased now-a-day’s due to the rapid growth of population in major cities. Overwhelming number of vehicles and insufficient roads are the major causes of traffic congestion. This needs new technologies to be adopted, and a better approach for effective traffic management. In the literature, researchers use conventional methods such as IR sensor, wireless sensor, and Fuzzy logic to measure the traffic density. The main limitations of such conventional methods are that they require personal monitoring of the traffic and ineffective to work in foggy weather. The main aim of this work is to develop a real-time adaptive density-based traffic management system that can quantify number of vehicles on roads under foggy weather conditions. The proposed system involves video acquisition, frame extraction, fog removal and vehicle counting. At first, the video is captured by camera and split into number of frames using frame extraction process. The Dark channel prior (DCP) algorithm is used to remove the fog from each frame and the background subtraction method and certain morphological operations are used to count the number of vehicles in real-time. Based on the vehicle count, the system specifies the time required to clear the traffic. This could facilitate ease traffic flow, save time, and even operate in foggy weather conditions, which is an improvement from the conventional timer-based operations of traffic signals. Key words : FOG, Haze, Image, DCP, Airlight, Atmospheric Light, Vehicle. 1. INTRODUCTION Now a day the utilization of automobiles is increasing every day which is causing more traffic congestion. The traffic congestion is a significant problem in Urban and Metro Politian areas. Due to increase in population in these areas causes to increase in vehicular density, traffic related problems. These problems leads to delayed services increase in transportation costs and fuel consumption. Due to the using of conventional traffic light systems at most of the traffic junctions, which are working based on fixed time concept leads to clumsy traffic flow. Figure 1: Traffic image in Haze condition For example if we consider two lines where one is having more vehicular density and another is having less vehicular density but both got same time of green signal where it is loss of time for one line. By considering this problem if we provide signal based on the vehicle density, we can avoid the loss of time and clumsy traffic flow. In the automatic light systems video cameras are used to record traffic information continuously. We can analyze traffic videos to detection and tracking of vehicles for traffic flow analysis by using different methods like Mixture of Gaussian (MOG), Code Book (CB) and Modified Code Book (MCB) [1], Tracking learning Detection (TLD) [2], Struck: Structured Output Tracking with Kernels [3] and High-Speed Tracking with Kernelized Correlation Filters [4] to count the number of vehicles passed per given time [5,6]. Density based Traffic Monitoring system in Haze and FOG Conditions G. HARISH BABU 1,2 , N. VENKATRAM 3 , 1 Koneru Lakshmaiah Education Foundation, India, harish.sidhu12@gmail.com 2 CVR College of Engineering, India, harish.sidhu12@gmail.com 3 Koneru Lakshmaiah Education Foundation, India, venkatram@kluniversity.in ISSN 2347 - 3983 Volume 9. No. 3, March 2021 International Journal of Emerging Trends in Engineering Research Available Online at http://www.warse.org/IJETER/static/pdf/file/ijeter22932021.pdf https://doi.org/10.30534/ijeter/2021/22932021