基于视觉技术的煤块检测方法设计  

Design of Coal Briquette Detection Method Based on Vision Technology

作　　者：赵奇 ZHAO Qi(Guoneng Nortel Shengli Energy Co.,Ltd.,Xilingol 026000,China)

机构地区：[1]国能北电胜利能源有限公司,内蒙古锡林郭勒026000

出　　处：《自动化仪表》2024年第8期122-126,共5页Process Automation Instrumentation

摘　　要：为了解决选煤过程中效率低、污染性大和投入力度大等问题,并提高选煤的质量和效率,提出一种融入Roberts算法和阈值式的X射线煤块检测方法。首先,介绍了机器视觉、X射线、图像分割等煤块检测技术。基于以上技术,设计了煤块和矸石分选阈值的方法。然后,通过试验,测试了煤块和矸石的参数、峰值灰度级和厚度、煤块阈值和矸石阈值,以及不同厚度下系统对煤块的识别精度。最后,对比了传统识别系统所需设备与基于Roberts算子和X射线的识别系统所需设备的成本。试验结果表明,所提方法能够减少物体厚度对识别结果产生的负面影响,从而进一步提高识别精度;能够较好地识别和区分煤块和矸石,且性价比较高。该方法对于煤块检测的识别具有重要的参考价值,能在一定程度上推动图像识别领域和能源利用领域的技术发展。To solve the problems of low efficiency,high pollutability and high input efforts in the process of coal selection,and to improve the quality and efficiency of coal selection,an X-ray coal briqutte detection method incorporating Robert’s algorithm and thresholding is proposed.Firstly,the coal briqutte detection technologies such as machine vision,X-ray and image segmentation and so on are introduced.Based on the above technologies,the method of sorting thresholds for coal briqutte and gangue is designed.Then,through experiments,the parameters of coal briqutte and gangue,the peak gray level and thickness,the coal briqutte thresholds and gangue thresholds,as well as the system’s recognition accuracy of coal briqutte under different thicknesses,are tested.Finally,the cost of the equipment required for the traditional recognition system and the equipment based on the Roberts operator and X-ray recognition system are compared.The experimental results show that the proposed method can reduce the negative impact of object thickness on the recognition results,can further improve the recognition accuracy;can better recognize and distinguish between coal briqutte and gangue,and cost-effective.The method has an important reference value for the identification of coal briqutte detection and can promote the technological development in the field of image recognition and energy utilization to a certain extent.

关 键 词：机器视觉技术 煤块和矸石 X射线 ROBERTS算子 阈值理论 图像识别 

分 类 号：TH-3[机械工程]