Transportation with 29% share in overall greenhouse gas emissions, is a major source of urban health and environmental degradation. A lot of effort has gone into development of different solutions to analyse, control and manage traffic flow to reduce vehicular emissions. In this regard, a low-cost, easily installable and maintainable solution for traffic flow characterization is of utmost importance to provide true intelligent transportation solutions. In this work, a raspberry pi based cyber physical system has been proposed for vehicle counting using image processing. Moreover, the proposed solution has the capability to measure associated roadside vehicular emissions such as carbon dioxide, carbon monoxide and particulate matter. The proposed solution can be used to develop relationships between traffic flow and associated roadside pollutants. For data logging and analytics, the sensed parameters were transmitted to a free and open source cloud platform “ThingSpeak”. For field testing, the proposed solution was installed on a main thoroughfare in 42 minutes. Sensed parameters were transmitted per minute with 100% accuracy to ThingSpeak using Wi-Fi. Vehicle counting accuracy of the proposed system was 86.9%. On-road traffic flow was successfully characterized in terms of traffic flow, density and average time headway. Relationships between measured traffic flow parameters and associated sensed pollutants (carbon dioxide, carbon monoxide and particulate matter) were established. The proposed solution to the fabrication cost of $70 has the capability to operate for 13 hours without any human intervention.