摘要
受到复合采空区特征的影响,传统自燃风险图像识别方法识别误差较大,无法满足复合采空区煤自燃风险预警要求。为了彻底解决识别误差偏大问题,根据采空区漏风增加、遗煤氧化升温、氧浓度分布不均匀等特点,在图像识别处理上引入超像素分割方法与三维空间进行识别优化,具体可分为煤层复合采空区煤自燃风险图像预处理、煤层复合采空区自燃裂隙特征分析、自燃风险图像特征区域标记与煤层复合采空区煤自燃图像识别四部分。通过与其他识别方法的性能比较验证所提出的方法的有效性,其在现阶段内具备有效解决复合采空区煤自燃风险图像识别误差偏大问题的能力。
Due to the influence of the characteristics of composite goaf,the traditional image recognition methods for spontaneous combustion has large recognition errors and cannot meet the requirements for early warning of coal spontaneous combustion risk in composite goaf.In order to completely solve the problem of large recognition errors,according to the characteristics of increased air leakage in goaf,oxidation and temperature rise of residual coal,uneven distribution of oxygen concentration and so on,the super-pixel segmentation method and 3D space are introduced into image recognition processing to recognition optimization,which can be divided into four parts:image preprocessing of coal spontaneous combustion risk in coal seam composite goaf,analysis of characteristics of spontaneous combustion cracks in coal seam composite goaf,the feature area labeling of spontaneous combustion risk image,and the recognition of coal spontaneous combustion image in the composite goaf of coal seams.Compared with other identification methods,the proposed method is proved to be effective,and it has the ability to effectively solve the problem of large error in image recognition of coal spontaneous combustion risk image in composite goaf at present.
作者
任兴华
刘梁
宣金国
REN Xinghua;LIU Liang;XUAN Jinguo(Libi Coal Mine of China Coal Huajin Group Jincheng Energy Co.,Ltd.,Jincheng 048200,China;The Fifth Geological Brigade of Jiangxi Geological Bureau,Nanchang 330000,China;Anhui Wantai Geophysical Technology Co.,Ltd.,Hefei 230000,China)
出处
《微型电脑应用》
2025年第12期225-228,233,共5页
Microcomputer Applications
关键词
煤层复合采空区
煤自燃
超像素分割
图像识别
composite goaf of coal seams
coal spontaneous combustion
super pixel segmentation
image recognition
