Foreground Detection via Background Subtraction and Improved Three-Frame Differencing

Moving object detection is a widely used and important research topic in computer vision and video processing. Foreground aperture, ghosting and sudden illumination changes are the main problems in moving object detection. To consider the above problems, this work proposes two approaches: (i) improved three-frame difference method and (ii) combining background subtraction and improved three-frame difference method for the detection of multiple moving objects from indoor and outdoor real video dataset. This work accurately detects the moving objects with varying object size and number in different complex environments. We compute the detection error and processing time of two proposed as well as previously existing approaches. Experimental results and error rate analysis show that our methods detect the moving targets efficiently and effectively as compared to the traditional approaches.