基于LSM和RELBP的煤岩识别方法探讨
    点此下载全文
引用本文:张 荣 华.基于LSM和RELBP的煤岩识别方法探讨[J].计算技术与自动化,2021,(1):109-113
摘要点击次数: 73
全文下载次数: 0
作者单位
张 荣 华 (国家能源集团上榆泉煤矿山西 河曲 036500) 
中文摘要:对当前煤岩识别方法的研究现状进行了介绍,并提出将最小二乘法模型(Least square model ,LSM)和融入平滑滤波思想的鲁棒扩展局部二值模式(Robust extended local binary Pattern ,RELBP)融入煤岩识别领域。对基于LSM和RELBP的煤岩识别方法的煤岩自动化识别技术(RELBP-LSM)进行了探讨。结果表明:(1)当前的煤岩识别方法大多存在效果较差、稳定性欠佳、适用范围小等缺点,同时易受人为因素的影响;(2)以最小二乘法和局部二值模式为理论基础,建立起RELBP-LSM煤岩识别方法,并通过参数敏感性分析,确定正则化参数λ的最佳取值为10-3.5,优选模式数d的最佳取值为500;(3)对不同方法的准确识别率进行对比分析,认为RELBP-LSM法不仅具有较高的准确识别率,同时能大大降低内存占用率,加快识别速率和效率。
中文关键词:煤岩  最小二乘法  局部二值法  RELBP-LSM  参数敏感性  准确识别率
 
Discussion on Coal Rock Identification Method Based on LSM and RELBP
Abstract:This paper discusses the current research situation of coal and rock recognition methods, and puts forward that the least square model (LSM) and robust extended local binary pattern (RELBP), which are integrated into the idea of smooth filtering, are integrated into the field of coal and rock recognition, and the coal and rock automatic recognition technology (RElB) based on LSM and RELBP is applied to the field of coal and rock recognition P-LSM). The results show that: (1) most of the current methods of coal rock identification have some disadvantages, such as poor effect, poor stability and small application scope, and are easily affected by human factors; (2) based on the least square method and local binary model, the RELBP-LSM method of coal rock identification is established, and the best value of regularization parameter λ is determined to be 10-3.5 by parameter sensitivity analysis The best value of pattern number d is 500; (3) by comparing and analyzing the accuracy of different methods, it is considered that RELBP-LSM method not only has a high accuracy of recognition, but also can greatly reduce the memory occupation rate and speed up the recognition rate and efficiency.
keywords:coal rock  least square method  local binary method  RELBP-LSM  parameter sensitivity  accuracy recognition rate
查看全文   查看/发表评论   下载pdf阅读器