基于MATLAB的钨矿石粒度在线检测系统设计
来源:用户上传
作者:易畅 彭翠 熊微 王卫花 罗小燕
摘 要:针对目前钨矿石初选环节中选矿效率低、实时性和准确性差等问题,提出矿石粒度在线检测系统设计方案,通过MATLAB软件先对原始图像进行中值滤波处理,再分别采用维纳滤波、Lucy-Richardson和盲区卷积复原图像并对比复原效果,在考虑时间成本的情况下以Lucy-Richardson进行图像复原为最佳。考虑到矿石由不同形状组成,运用形态学处理的方法构造不同大小及形状的结构元素,通过开闭重建运算对图像进行腐蚀和膨胀,对孔洞填充及噪声处理。采用基于标记的分水岭分割算法,很好地解决了传统分水岭算法的过分割现象,提升了粘连矿石分割的准确率。最后通过矿石像素面积检测,统计矿石颗粒的粒径、粒级大小等相关信息,提高了矿石分选速度与准确度,满足工业化实时检测识别要求。
关键词:粒度检测;分水岭分割;形态学处理;MATLAB
中图分类号:TP391 文献标志码:A 文章编号:1003-5168(2022)2-0023-05
DOI:10.19968/j.cnki.hnkj.1003-5168.2022.02.005
Design of Online Particle Size Detection System for Tungsten Ore Based on MATLAB
YI Chang PENG Cui XIONG Wei WANG Weihua LUO Xiaoyan
(School of Mechanical and Electrical Engineering, Jiangxi University of Science and Technology, Ganzhou 341000,China)
Abstract: In view of the problems of low beneficiation efficiency, insufficient real-time and accuracy in the primary process of tungsten ore, a design scheme of online detection system for ore particle size was proposed. The original image was firstly median filtered by MATLAB software, and then wiener filtering, Lucy-Richardson and blind zone convolution were used to restore the image and compare the restoration effect. Under the consideration of time cost, lucy-Richardson is the best method for image restoration. Considering that ores are composed of different shapes, morphological processing method is used to construct structural elements of different sizes and shapes, and the image is corroded and expanded by opening and closing reconstruction operation, and the hole filling and noise processing are carried out. The watershed segmentation algorithm based on mark is adopted to solve the over-segmentation phenomenon of traditional watershed algorithm and improve the segmentation accuracy of cohesive ore. Finally, through the detection of the pixel area of the ore, the relevant information such as the particle size and size of the ore particle is counted, which improves the speed and accuracy of ore sorting and meets the requirements of industrial real-time detection and identification.
Keywords: particle size detection; watershed segmentation; morphological treatment; MATLAB
0 引言
u矿石是我国重要的矿产资源,目前钨矿石初选是通过纹理、形状大小、颜色等方面的判断,经过人工筛选出来[1]。这种传统的钨矿石初选方法的效率和准确性都不高[2]。而粒度检测可将采集图片的颗粒以不同等级大小的层度划分出来,并统计其数量,在各种生产工厂以及分析产品等领域发挥着极其重要的作用[3-4]。目前比较先进的检测有激光、电镜等,这些粒度检测的精度非常高,但是其对生产工作环境的要求颇高[5]。而在本研究的矿石粒度检测这一块显然是无法适用的,因为矿石破碎会产生大量的粉尘,并且高成本的仪器无法承担被损坏的风险[6]。粒度在矿石破碎中是一个非常重要的参数,其实时性和测量的准确性是最迫切需要解决的问题[7]。采用计算机图像处理技术,可以实现矿石粒度检测的自动化,成本低,并且识别速度快,准确率较高,为矿业生产自动化提供了很好的解决方案[8]。MATLAB是一个基于矩阵运算的、功能强大的软件,对于图像处理来说,每一张图像的像素信息可以看成一个大矩阵,因此MATLAB处理图像比较快速,可以大大节省矿石图像处理时间,即间接缩短整个流程中粒度在线检测的时间[9]。
转载注明来源:https://www.xzbu.com/1/view-15425494.htm