基于优化模糊集的视频图像火焰检测算法与实现  

VIDEO IMAGE FLAME DETECTION ALGORITHM BASED ON OPTIMISED FUZZY SETS AND ITS IMPLEMENTATION

作　　者：周一君[1] 马莉[1] 

机构地区：[1]杭州电子科技大学自动化学院,浙江杭州310018

出　　处：《计算机应用与软件》2011年第7期25-27,34,共4页Computer Applications and Software

基　　金：国家自然科学基金(60775016);浙江省重大科技专项(2007C13062)

摘　　要：视频监控系统的火焰自动检测成为公共安全领域关注的热点之一。根据图像亮度和色度特征,提出了一种基于优化模糊集的火焰检测算法。首先论述了火焰检测的模糊规则集建立,然后利用遗传算法离线优化输入/输出模糊集的隶属度函数,提高了模糊推理控制精度。并采用TMS320DM642视频图像处理硬件平台,实验结果表明模糊推理和遗传算法结合的火焰检测具有较高的检测精度。Automatic flame detection technology of the video surveillance system has become one of the hot topics in the public security area.On the basis of such image features as luminance and chrominance,a flame detection algorithm based on optimised fuzzy sets is proposed.This paper first discusses the establishment of fuzzy rule sets of flame detection,then utilises the genetic algorithm to offline optimise member functions for input/output fuzzy sets,so that the precision of fuzzy deduction control is improved.TMS320DM462 video image processing hardware platform is adopted.Experiments show that the flame detection algorithm combining both fuzzy deduction and genetic algorithm can achieve higher detection accuracy.

关 键 词：TMS320DM642 模糊推理 遗传算法 隶属度函数 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]