基于健全逻辑的煤矿监测预警决策模型研究  

Monitoring and Early Warning Decision Model of Coal Mine Based on Sound Logic

作　　者：陈佳林 CHEN Jialin(National Institute for Occupational Safety,China Coal Information Institute,Beijing 100029)

机构地区：[1]国家安全生产监督管理总局信息研究院,北京100029

出　　处：《计算机与数字工程》2017年第6期1138-1140,1193,共4页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:61302163)资助

摘　　要：煤矿监测图像的准确识别对于预警决策意义重大。由于模糊推理方法的选择很大程度地影响推理的性能和效果,基于健全逻辑建立了交互信息与广义相关系数的相互关系,改进了模糊决策树选择匹配算子的算法。结合煤矿监测图像数据特点,为了获取最佳特征组合,根据图树转换规则,提出了多目标部位提取特征的方法。通过对煤矿监测图像数据的检验分析,对比传统的针对整张图像或感兴趣区域进行特征提取的方法,该方法能获得更高的预警决策正确率。In the practical application of fuzzy decision tree in coal mine,model of matching operator in the fuzzy decisiontree is improved about the relationship of mutual information and generalized correlative coefficient based on sound logic.In order toget the best combination of the features,the algorithm that extracts the shape feature considering the adjacent parts is proposedbased on the idea of chart transferring tree in the tree structure comprehensively.Compared with the traditional extracting algorithmswhich aim at the whole picture or a single part,the experiments by examination and analysis of clinical cases show that the proposedalgorithm can obtain a higher early warning accuracy.

关 键 词：安全监测 健全逻辑 模糊决策树 交互信息 特征提取 

分 类 号：TD76[矿业工程—矿井通风与安全]