An Automated Road Traffic Congestion Estimating Program

Video processing techniques have been used in determining road traffic conditions and it proved itself quite useful but still cannot adapt to changes in the environment lighting. Human intervention is required to input the new background which would require effort. The project focuses on the creation of a program that uses background elimination in extracting the foreground objects. However, to solve the problems regarding change in lighting, the background estimating algorithm must be adaptive to this change. It uses the Kalman tracking equations to derive the mean update and standard deviation using tuned values of measurement noise covariance and process noise covariance. The background to be subtracted will then update itself depending on the changes in lighting during the video stream and any slight movement of the camera itself. Blob analysis is then used to gather data. Total blob area and centroid movement are the parameters to be considered in the program's decision making and fuzzy logic is used to decide given these parameters. The output is displayed as an embedded text in the original video. The whole program is created and simulated in the MatLab environment using its "vision" package.