Detection of Moving Vehicle using Adaptive Threshold Algorithm in Varied Lighting

Designing an accurate tracking algorithm of vehicles captured by a camera, which move from frame to frame in a video sequence is continuing to leave substantial obstacles. The biggest challenges come from different conditions of significant lighting changes, shifting positions of moving vehicles, vehicle's non-linear deformations, noise gained in data retrieval as well as vastly switching backgrounds. Vehicle tracking from videos recorded at night has a higher difficulty level than the one at daytime. The lighting changes, particularly at night, produce a very low-quality video recording and the resulting image. The reason is that the intensities of lighting at night often change rapidly and drastically. Background subtraction method is frequently used in solving vehicle tracking problems. Nevertheless, it has a weakness which gives noise or disturbance effects. By proposing the adaptive threshold algorithm derived from the Fuzzy C-Means (FCM) algorithm in this study, the accuracy of detecting moving vehicles in minimal lighting can be improved. The results of this algorithm are examined by using Mean Square Error (MSE) and Peak Signal Noise Ratio (PSNR) parameters.