Detection of Moving Object Using Morphological Filters

In this paper, novel morphological filters are developed under the scope of traffic system in India is proposed. The algorithms of three filters are developed and implemented with their proper coding using Matlab (R2017a) software to detect the moving objects from CCTV video signal. For this aim, three filters are designed with gaining concepts of linear filters and also non-linear operators i.e., morphological operators. Noise reducing is also important to identify or detect a moving object. As the most of traffic videos contain background images and also different noise signals, it is necessary to minimize or to eliminate noise by subtracting background images from the images of traffic video. After detecting moving object using three morphological filters developed, PSNR and SNR values are also calculated for identified object to get the best filter designed. It is seen from the result, that moving object i.e., only white car detected after removing noise and applying median filter followed by morphological filter on background subtracted image, gives highest PSNR and SNR values.