基于YOLOv3的煤矸识别系统  

Coal Gangue Recognition System Based on YOLOv3

作　　者：郑施航 陆金桂[1] Zheng Shihang;Lu Jingui(School of Mechanical and Power Engineering,Nanjing Tech University,Nanjing 210000,China)

机构地区：[1]南京工业大学机械与动力工程学院,南京210000

出　　处：《煤矿机械》2023年第5期187-189,共3页Coal Mine Machinery

摘　　要：对于传统图像处理算法在煤矸识别领域运用时存在的识别准确率低的问题,提出了一种基于YOLOv3的识别系统,该系统先对图像进行图像锐化等预处理操作,再将处理后的图像作为YOLOv3算法的输入,最终输出煤矸识别定位结果。实验表明,图像预处理结合YOLOv3的识别正确率达到了95.4%,相比于传统算法具有一定的优势。To solve the problem of low recognition accuracy in the application of traditional image processing algorithms in the field of coal gangue recognition,a recognition system based on YOLOv3 is proposed.The system first preprocesses the image,then uses the processed image as the input of YOLOv3 algorithm,and finally outputs the coal gangue recognition and positioning results.Experiments show that the recognition accuracy of image preprocessing combined with YOLOv3 reaches 95.4%,which has certain advantages over traditional algorithms..

关 键 词：图像处理 YOLOv3 图像锐化 煤矸识别 

分 类 号：TD94[矿业工程—选矿]