一种基于FPGA的快速视频背景分割方法  

A Fast Method for Reconstruction of Background Based on FPGA

作　　者：贾晓岚[1] 朱尤攀[2] 赵薇薇[2] 邹瑞滨[2] 彭廷海[2] 赵江来[2] 

机构地区：[1]郑州航空工业管理学院计算机系 [2]昆明物理研究所,昆明650223

出　　处：《红外技术》2007年第11期675-677,共3页Infrared Technology

摘　　要：背景的有效分割是实现动目标检测和识别的基础。针对目前众多算法在进行分割时面临的算法复杂度高和占用存储空间过大的问题,在对监视视频序列研究的基础上,使用低成本的Spartan3系列的FPGA平台,以小于16 Mbit的存储空间,实现了对视频序列静态场景即时重建。Effective background segmentation is the basement to accomplish the detection and classification of motion targets in many video applications. Many algorithms that can complete the reconstruction of background have the complicated structures, which can not be accomplished in a short time, and the exhausted occupation of memory space also disturbed the implementation of most algorithms. According to research of video surveillance, the essay puts forward to a method based on the low-priced Spartan3 platform whose memory is less than 16Mbit, which is proved that the result of reconstruction can keep a stable refresh in less than 1 second.

关 键 词：视频分割 静态场景 重建 现场可编程门阵列 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]