• 论文
主办单位:煤炭科学研究总院有限公司、中国煤炭学会学术期刊工作委员会
基于深度特征的煤矿冲击地压和煤与瓦斯突出图像感知报警方法
  • Title

    Coal mine rock burst and coal and gas outburst image perception alarm method based on depth characteristics

  • 作者

    程继杰刘毅

  • Author

    CHENG Jijie;LIU Yi

  • 单位

    中国矿业大学(北京)人工智能学院

  • Organization
    School of Artificial Intelligence, China University of Mining and Technology-Beijing
  • 摘要

    冲击地压和煤与瓦斯突出容易造成煤矿重特大事故。针对目前煤矿冲击地压和煤与瓦斯突出事故发生主要靠人工发现,结合灾害造成不同于正常工况下采掘工作面和巷道空间的颜色和深度特征,提出了一种基于深度特征的煤矿冲击地压和煤与瓦斯突出图像感知报警方法:在煤矿井下巷道顶板、巷帮靠近顶板以及采掘工作面液压支架顶部或靠近顶部等位置多点布置带有补光灯的彩色双目摄像机,实时采集采掘工作面和巷道空间彩色图像和深度图像;以煤矿井下与灾害抛出煤岩有着鲜明颜色差别的生产设备作为背景,监测识别是否出现彩色图像颜色发生较大变化;如果掘进巷道端头、巷道中间、巷道入口,或回采工作面,或进风巷道入口、巷道中间,或回风巷道入口、巷道中间,或主运输、辅助运输大巷等位置彩色双目摄像机监测到图像颜色发生较大变化,则监测彩色图像平均亮度是否小于设定的亮度阈值;如果平均亮度小于设定的亮度阈值,则以与灾害抛出煤岩有鲜明颜色差别的生产设备为背景,监测深度图像是否发生较大变化;如果深度图像发生较大变化,那么监测导致深度图像发生较大变化物体的移动速度是否大于设定的速度阈值(V>13 m/s);当移动速度大于设定速度阈值时,利用多点布置的甲烷传感器监测监视区域内甲烷浓度,当采煤工作面、掘进工作面、进风巷道、回风巷道、总回风巷道等多个不同地点均监测到甲烷浓度大幅升高或达到报警值,那么进行煤与瓦斯突出报警,反之,则进行冲击地压报警。提出了彩色双目摄像机感知灾害最佳倾角的确定方法,以及导致深度图像发生较大变化物体移动速度的确定方法。在确保安全性的前提下,简化了灾害模拟试验,设计了一套煤矿冲击地压和煤与瓦斯突出颜色和深度特征模拟试验装置:采用与煤岩颜色和比重相近的10 mm橡胶球替代灾害抛出煤岩;利用直径315 mmPVC管模拟煤矿井下巷道受限空间以及与抛出煤岩有着鲜明颜色差异的背景设备;利用轴流式高压鼓风机作为动力装置,进行模拟冲击地压和煤与瓦斯突出发生时大量煤岩抛出造成的采掘工作面和巷道空间颜色和深度异常特征;利用3.4 mm焦距、30FPS、71°×55°视场角的彩色双目摄像机,完成对整个灾害模拟过程的监视与彩色双目图像采集。研究分析表明,提出的方法可识别出灾害发生时颜色和深度特征,验证了基于深度特征的煤矿冲击地压和煤与瓦斯突出图像感知报警方法的可行性与有效性。

  • Abstract

    Rock burst and coal and gas outburst are easy to cause serious accidents in coal mines. In view of the current coal mine rock burst and coal and gas outburst accidents occurring by manual discovery, combined with the color and depth characteristics of mining face and roadway space cause by the disaster that are different from normal working conditions, a coal mine rock burst and coal and gas outburst image perception alarm method based on depth characteristics is proposed. Firstly, visible light binocular cameras with supplementary light lamps are arranged at multiple points in the roof of the roadway, near the roof of the laneway's side and in or near the top of the hydraulic support of the mining face to collect real-time color and depth images of working face and roadway; Secondly, the production equipment in coal mines with distinct color differences to the coal rocks thrown out by disasters is used as the background equipment to monitor and identify whether there is a significant change in the color of the color image; Third, when the end, the middle and the entrance of the driving roadway, or the mining face, or the entrance and the middle of the air inlet roadway, or the entrance and the middle of the return air roadway, or the main transportation and auxiliary transportation roadway and other positions of the visible light binocular camera to monitor the image color changes greatly, then monitor whether the average brightness of the color image is less than the set brightness threshold; Fourthly, when the average brightness is less than the set brightness threshold, the depth image is monitored with the production equipment that has a distinct color difference from the coal rocks thrown by the disaster as the background. Fifth, when the depth image changes greatly, whether the moving speed of the object resulting in a large change in the depth image is greater than the set speed threshold(V>13 m/s) is monitored; Sixth, When the moving speed is greater than the set speed threshold, methane sensors arranged at multiple points are used to monitor the methane concentration in the monitoring area, and when the methane concentration of coal mining face, driving face, inlet roadway, return roadway, total return roadway and other different locations were detected increased significantly or reached the alarm value, then the coal and gas outburst alarm is carried out, otherwise, the rock burst alarm is carried out. A method to determine the optimum dip angle of visible light binocular camera is proposed, and a method to determine the moving speed of objects that cause great changes in depth image is proposed. On the premise of ensuring security, simplifying the disaster simulation experiment, a set of experimental equipment to simulate the color and depth characteristics of rock burst and coal and gas outburst is designed: the 10 mm rubber balls with similar color and specific gravity of coal rock is used to replace the disaster throwing coal rocks; and the 315 mmPVC tubes is used to simulate the confined space of the roadway and the background equipment with distinct color difference from the coal rock thrown from the coal mine; and the axial flow high pressure blower is used as the power device to simulate the abnormal characteristics of color and depth of mining face and roadway caused by a large number of coal and rock during the occurrence of rock burst and coal and gas outburst; and the visible binocular camera with focal length of 3.4 mm, 30FPS and field of view angle of 71°×55° is used to complete the whole disaster simulation process monitoring and visible binocular image acquisition. Research and analysis show that the proposed method can identify the color and depth characteristics when the disaster occurs, and verify the feasibility and effectiveness of the image perception alarm method of rock burst and coal and gas outburst based on depth characteristics.

  • 关键词

    冲击地压煤与瓦斯突出颜色特征深度特征速度特征图像识别

  • KeyWords

    rock burst;coal and gas outburst;color characteristic;depth characteristic;velocity characteristic;image recognition

  • 基金项目(Foundation)
    国家重点研发计划资助项目(2016YFC0801800)
  • DOI
  • 引用格式
    程继杰,刘 毅. 基于深度特征的煤矿冲击地压和煤与瓦斯突出图像感知报警方法[J]. 煤炭科学技术,2024,52(3):245−257.
  • Citation
    CHENG Jijie,LIU Yi. Coal mine rock burst and coal and gas outburst image perception alarm method based on depth characteristics[J]. Coal Science and Technology,2024,52(3):245−257.
  • 相关文章
  • 相关专题
  • 图表
    •  
    •  
    • 煤矿井下典型设备颜色特征

    图(10) / 表(0)

相关问题

主办单位:煤炭科学研究总院有限公司 中国煤炭学会学术期刊工作委员会

©版权所有2015 煤炭科学研究总院有限公司 地址:北京市朝阳区和平里青年沟东路煤炭大厦 邮编:100013
京ICP备05086979号-16  技术支持:云智互联