Fast denoising for moving object detection by an extended structural fitness algorithm
暂无分享,去创建一个
Aim of the paper is to propose an optimized method to remove from the images taken by a camera the most common noises, i.e., the ones due to "waving trees" or to bad digitalization. This makes a real time identification of the objects traversing windy scenes or scenes taken with cameras affected by salt and pepper noise possible. The method is based on an extension of the Structural Fitness (SF) algorithm to avoid the analysis of those pixels that can be safely considered to belong to the background according to a simple statistical formula proposed in the paper. The experiments discussed in the paper demonstrate how the denoising method works and its excellent time performance and accuracy in typical engineering applications.
[1] Luigi di Stefano,et al. Analysis of pixel-level algorithms for video surveillance applications , 2001, Proceedings 11th International Conference on Image Analysis and Processing.
[2] Concetto Spampinato,et al. Soft-Computing Agents Processing Webcam Images to Optimize Metropolitan Traffic Systems , 2004, ICCVG.
[3] Alessandro Bevilacqua,et al. Robust denoising and moving shadows detection in traffic scenes , 2001 .