论文部分内容阅读
在实时嵌入式视频监控系统中,对于一个高效率的移动侦测算法研究是一个热门课题。本文提出一个基于16路视频采集卡报警系统的移动侦测算法。在不降低其精确度的情况下,能比较明显的提高侦测效率。采用对背景采样的渐变算法,降低背景快速变化时CPU的大量计算;利用阈值取法的自适应性算法,去除了复杂的数学模型带来的繁杂的计算,从而减少CPU的负荷;本算法能适应于不同场所、不同灵敏度的侦测要求;论文给出了本算法在实时侦测时的分块模式和实验结果。
In real-time embedded video surveillance system, research on a highly efficient motion detection algorithm is a hot topic. This paper presents a 16-channel video capture card alarm system motion detection algorithm. Without reducing the accuracy of the case, can significantly improve the detection efficiency. The gradient algorithm of background sampling is adopted to reduce the massive computation of CPU when the background is rapidly changing. The adaptive algorithm of threshold extraction removes the complicated calculation caused by the complex mathematical model, thus reducing the CPU load. The algorithm can adapt to In different places, different sensitivity detection requirements; the paper gives the algorithm in real-time detection of the block mode and experimental results.