基于机器视觉的煤炭自动识别分拣系统研究

摘要

In order to solve the problems of labor-intensive, low efficiency and accuracy in recognition and sorting of coal, designed an automatic coal recognition and sorting system based on machine vision which consisted of hardware and software. In order to achieve the automatic identification of coal and gangue, a new method based on gray-level co-occurrence and texture analysis was proposed, realized the automatic recognition through training and classifying the images of coal and gangue with SVM , programmed the image processing software and system interface software. The test results show that the system hardware selection is rational , and the system software is highly stable and efficient which improved the efficiency and accuracy in recognition and sorting of coal with practical significance and high economic value. Meanwhile, the system can also be extended to sorting of other materials.%为了解决煤炭人工识别分拣劳动强度大,效率和精度低的问题.基于机器视觉技术,设计了一套由硬件系统和软件系统组成的煤炭自动识别分拣系统.为实现煤炭和矸石的自动识别,提出了基于灰度共生矩阵的纹理分析识别算法,并通过支持向量机(SVM)对煤炭图像和矸石图像样本进行训练和分类从而实现自动识别功能,并完成图像处理算法与系统界面软件编程.实验结果表明,该系统硬件选型合理,算法稳定性高,效率高,提高了煤炭识别分拣的效率与精度,具有很高的实际意义和经济价值.同时,该系统也可推广到其他物料的自动识别分拣中.