摘要
研究将贝叶斯决策应用于自适应神经-模糊推理系统(ANFIS)的视频烟雾检测系统。提取视频烟雾特征,通过减法聚类和混合学习算法,确定并优化得到ANFIS实例,引入贝叶斯决策对ANFIS输出进行检测判别。仿真实验表明,ANFIS比其他烟雾检测算法具备更好的检测性能,而基于最小风险的贝叶斯决策可进一步提高检测率和降低虚警率,能更好地满足实际应用的需求。
The Bayesian decision method is studied to further improve the performance of detecting video smoke using Adaptive Neuro-Fuzzy Inference System（ANFIS）. Smoke features are extracted from video sequences. The subtractive clustering and hybrid learning rules are used to train ANFIS. Detection outputs are determined by performing proposed Bayesian decision rules on the outputs of ANFIS. Experimental results show that the detection performance of ANFIS is better than that of other smoke detection algorithms, and the introduction of minimum risk-based Bayesian decision rules further increases the detection rate and decreases the false alarm rate, which is more valuable for practical applications.
出处
《计算机工程与应用》
CSCD
2014年第3期173-176,共4页
Computer Engineering and Applications
基金
2011年度江南大学自主科研计划项目(No.JUSRP11125)
