earticle

영어

Recently, due to the interest for personal safety, intelligent image recognition technology using the CCTV has received a lot of attention in many areas. Real-time video or image tracking technology is the technology to continuously track the movement of the target object and maintain the specified information of the object. However, the problem is that unnecessary noise or environmental changes of the input video from CCTV make it difficult to accurately detect and track the target object. In this paper, we propose a fast object tracking algorithm, which is divided into two parts: i) object detection process and ii) object tracking process. In the background image replacement technique, we compute the difference between the initial input image and input video image. Then, it replaces N×M blocks based on the changed pixels, finally leading to replacing the background images. If it changes over the pre-defined threshold value, we set the block size and generate the variable search window for object tracking. Since the variable search window only generates the minimum search area for object tracking, our proposed method is quickly able to compute the variable region for object tracking. Through various experiments, our proposed scheme outperforms the other object detection and tracking methods in terms of object tracking counts and tracking error counts.