Object detection, tracking and counting using enhanced BMA on static background videos

The availability of high quality and inexpensive video cameras, and the demand for automated video analysis has generated a great deal of interest in object tracking, counting algorithms. This paper presents an approach to count the moving objects or vehicles of traffic scenes recorded by static cameras. Different algorithms and sensors are used for object detection tracking and counting which increases the overall cost and gives less accurate result due to intensity of camera used for recording the video traffic. This paper proposes enhanced BMA(Block matching algorithm) with counting by using kernel tracking (template matching). Background subtraction technique is used to extract moving object from videos subsequently, then BMA and dilution technique is used to isolate and identify image blocks as single vehicle or object. Choice is given to user for selecting one or two different region to count the object. If the object passes from the region then only it will be counted. Finally the number of count of objects is done and sum of all objects (vehicles) is calculated in case of multiple lanes.