Formation and instability of spiral wave induced by Gaussian coloured noise

In this paper, we studied the effect of Gaussian coloured noise on the formation and instability of spiral waves described by one class of modified FitzHugh Nagumo equation. It was found that Gaussian coloured noise plays a constructive role in the formation, transition and instability of spiral wave. Too weak or too strong noise may act against the formation of spiral waves. At a certain noise level, spiral wave is maintained in a medium, in which spiral wave cannot be observed in the absence of the noise. It is difficult to make a stable spiral wave into unstable state by Gaussian coloured noise, unless the noise level is very high. The parameter regions of Gaussian coloured noise for spiral forming and spiral instability were given and discussed with numerical simulations.

结合LBP直方图和SVM的视频火焰检测

根据火焰独特的纹理特征,提出以块的LBP直方图特征为主并结合其他动、静态特征的火焰识别算法。首先用帧差法和RGB颜色高斯模型进行运动颜色检测,得到疑似火焰区域;再提取其红色分量统计特征、小波高频能量和LBP直方图特征;最后将特征向量输入SVM分类器进行火焰识别。实验证明,该算法误报率低、鲁棒性强,同时具有实时性,火焰视频检测率可达到96.2%。