Adaptive thresholding and Wagon counting technique for day and night time images

A novel approach for train wagon counting during day, afternoon, evening and night time is proposed in this paper. First step of the algorithm is to determine whether the captured video belongs to the day time or to the night time. This is done by plotting the histogram of the background reference frame. Next step is to find the adaptive threshold according to the illumination condition of the video captured. The output of the adaptive thresholding block is given as an input to the segmentation algorithm for foreground object detection. Different from traditional segmentation algorithms, this method deals with only the central reference mask pixel values. This increases the processing speed of the algorithm as well as efficiency of the system. The pixel difference values from the central reference mask, determine the segmentation of background and the foreground objects, which here, in this case are railway wagons.