摘要
提出一种基于星状骨架和隐马尔科夫模型的行走姿态识别方法,以猪为识别对象,采用猪体星状骨架模型提取特征向量,将特征向量进行符号映射,实现了将时间序列的姿态影像转换成符号序列,并对训练得到的隐马尔科夫模型进行测试。实验结果表明,该方法对猪的3种行走姿态有较好的识别效果。
An approach for pig walking posture recognition based on star skeleton and hidden markov model is presented. Extract feature vectors by using pig star skeleton. We realized time-sequential posture images transforming into symbols sequence by mapping features to symbols. And we test the trained HMMs, the experimental results show that this meth- od has a good recognition for pig＇s three walking posture.
出处
《软件导刊》
2015年第10期47-50,共4页
Software Guide
基金
江苏省自然科学基金资助项目(BK20140467)
江苏省高校自然科学研究面上项目(13KJB210006)
关键词
图像处理
模式识别
星状骨架
隐马尔科夫模型
姿态识别
Image Processing
Pattern Recognition
Star Skeleton
Hidden Markov Model
Posture Recognition
