基于低秩矩阵恢复交通视频背景重建性能评价研究.pdfVIP

M ay COM PU T ER ENGINEERING A ND DESIGN Vol No SBM I T P A doi issn j Perf o rmance evaluation of traff ic video backg round reconstruction based on low rank matrix recovery C HEN Chuan JI Xiaoia C H EN Zhe j Naval Aeronautical Engineering Academics Qingdao Branch Qingdao China School of Information Engineering Chang an U niversity Xian China Abstract For evaluating the performance of low rank matrix recovery LRM R algorithms in the background reconstruction of traffic video a series of tests w ere performed on matrix completion M C robust principal component analysis RPCA and low rank representation LRR to recover the traffic video background T hese algorithms w ere compared w ith some traditional background recovery methods including w eighted moving average adaptive background learning and Gaussian mix ture model In the ex periments three situations w ere taken into account ie illumination changing different traffic flow and shadow A group of metrics recommended by SBM I w as adopted as evaluation criterions as w ell as the common algorithm performance metrics such as CPU time and algorithm efficiency Ex perimental results show that for the slow change traffic scene LRM R al gorithms outperform the traditional methods in over