摘要
为预防煤仓满仓或空仓时引起安全事故,实时准确获取煤仓煤位信息,实现煤仓煤位可视化,研发了一套煤仓煤位监控系统,基于边缘检测法以及双目成像法,对双目视图配准的煤位图片进行处理,从建立的多尺度空间中提取煤位信息图片的特征点,通过对比左右煤位视图,利用双目成像视图配准的策略,进行深度匹配,解决了煤位图像处理和煤位煤仓煤位计算等问题,实现了对煤仓煤位的自动检测和控制。后期现场测试表明,软件预测煤位与实际煤位最大误差率不超过1.8%,证实了此套方法检测精度高,抗干扰性能强,能有效解决煤仓煤位的检测问题,具有一定的推广价值。
In order to prevent safety accidents caused by overwhelmed or empty storage,obtain real-time and accurate information of coal storage,and realize the visualization of coal level in coal bunker,a monitoring system for coal level of coal bunker is designed.The image of binary-view coal level is processed,and image feature points of coal level information are extracted from a multi-scale space.Through comparing the coal level image of right and left view,the images are deeply matched to solve the problems in coal level image processing and coal level computation with binocular imaging view registration strategy,thus to automatically detect and control the storage level.In the field test,the system showed high detection accuracy,with a maximum error rate of 1.8%,and strong anti-interference performance.
作者
赵睿
丁志刚
冯俊优
郭雄伟
ZHAO-Rui;DING Zhi-gang;FENG Jun-you;GUO Xiong-wei(Shenhua Group Zhungeer Energy Co.,Ltd.,Ordos 010300,China)
出处
《煤炭工程》
北大核心
2020年第S02期14-18,共5页
Coal Engineering
关键词
煤位监控
图像处理
煤位计算
双目视觉
coal level monitoring
image processing
coal level computation
binocular vision
