摘要
针对传统的闸机传感器识别系统在检测闸机口跳闸事件的不足,文中提出了一种基于图像处理的闸机口跳闸事件检测方法;该方法利用模板匹配的方法进行预处理,得到闸门图像匹配率;其次,根据跳闸事件的连贯性特点,提取一段连续时间序列的匹配率作为分类特征;最后,采取基于最小错误率的贝叶斯分类方法对这段时间序列特征向量进行分类;实验结果表明,该算法可以有效地检测出闸机口的跳闸事件,具有实时性好、不需要额外的传感器、成本低的优点,具有较好的工程应用前景。
Due to the shortcomings of the traditional gates sensor recognition systems in detecting jumped-out of gates,this paper proposes detection method based on image processing.Firstly,by using a template matching method to preprocess,the image matching rate of the gate can be obtained; Secondly,according to the continuity of the jumping out of gates,the feature vector extracted in this paper is time series of matching rate of the gate.Followed by is to classify the extracted feature vector by adopting the minimum error rate Bayesian classification.Experimental results show that the algorithm can effectively detect the jumping out of gates with the advantages of good real-time performance,no additional sensors and low costs.Therefore,this technology has a good prospect in engineering application.
出处
《计算机测量与控制》
2015年第1期164-166,共3页
Computer Measurement &Control
关键词
跳闸
特征提取
模板匹配
贝叶斯分类
jumped out of gates
feature extraction
template matching
Bayesian classification
