The project supported by Tubitak, which is carried out by Kuzu and his team in ITU, aims to immediately predict the air pollution caused by vehicles by bringing the dynamics of artificial intelligence and the calculation of liquids together. The developed method recognizes the density of the vehicle by traffic cameras and creates high -resolution air quality cards with this data and provides more precise predictions about the air pollution levels in cities.
Lamb, traffic, industry and domestic warming in cities as three large pollutants and traffic significantly influence air quality, he said.
The lamb observed carbon monoxide, particle substances and nitrogen oxides in all capitals, the emissions, and various difficulties in monitoring pollutant elements were occurred.
Easy access to data for modeling and forecasting is important that existing data is mainly based on general or average values, and it is not possible to access detailed data that is specific for each region. he said.
How does the system work?
Kuzu said that the systemic function of the method consists of three basic stages and is continued as follows:
“Firstly, the images preserved by traffic cameras are analyzed with deep learning algorithms and the vehicles are classified and their speeds are determined, each vehicle is determined. In the second stage, the amount of emissions of each vehicle group is calculated by using the emission factors that are specific for the type of vehicle types.
Kuzu explained that you can use the method that you have developed if you want to estimate emissions in the main street or on the side street everywhere and that it is possible to receive concentration data that is required for various reasons.
The accuracy of the results was tested
Kuzu emphasized that they use traffic monitoring cameras in Istanbul for the project and that this model was trained under different meteorological conditions and said that they could apply this model to the traffic cameras in Istanbul.
In the pilot study, Kuzu explained that traffic cameras were used in Besiktas Barbaros Boulevard.
“As a result, we can predict the vehicles over 95 percent. The calculation method is created with emissions and we can calculate the concentration at any point with a video image. The model was used for the air quality measuring station at the borders of the Besiktas Campus of the Technical University of Yıldız and the results have been tested.”
In the software you have developed, the main point is to determine the number and type of vehicles, and for this purpose, the information that the image is sufficient to process Kuzu has completed its words as follows:
“Firstly, we already know that the main emission source in our cities is traffic. There are two main pollutants in which the concentration limit is exceeded in cities. Can be processed.”
Bir yanıt yazın