Intelligent Post-processing via Bounding-Box-Based Morphological Operations for Moving Objects Detection

The detection of moving objects is a critical first step in video surveillance. Numerous background subtraction, frame differencing, optical flow algorithms and a number of post-processing techniques (including noise removal, binary morphological operations, and area thresholding) are used to extract the moving objects. However, these post-processing methods are time consuming and inefficient in real-time applications; for example, noise removal and binary morphological operations require scanning the video frame many times. The study presents an innovative post-processing technique, using bounding-box-based morphological operations, for grouping concentrated connected components and the removal of spread and small connected components for moving objects detection. Results demonstrate that the proposed method is more effective and efficient than traditional post-processing methods.