摘要
针对传统的煤矸石检测方式成本较高、识别准确率较低、适用性较差等不足,经分析实际检测要求,设计了一种基于机器视觉和AlexNet网络的煤矸石检测系统。该系统通过工业相机来采集传送带上煤矸石图像,利用直方图均衡化和二阶微分线性算子来加强图像对比度与锐化效果,并使用高斯滤波来抑制图像噪声,进而获取更具辨识度的图像,最终运用AlexNet网络实现煤矸石的识别与定位。结果表明,该系统识别准确率达到了95.90%,准确率较高,且实现过程较为简单,适用性良好。
Aiming at the shortcomings of traditional coal gangue detection methods, such as high cost,low recognition accuracy and poor applicability, after analyzing the actual detection requirements, a coal gangue detection system based on machine vision and AlexNet network is designed. The system collects the coal gangue image on the conveyor belt through the industrial camera, uses histogram equalization and second-order differential linear operator to strengthen the image contrast and sharpening effect, and uses Gaussian filter to suppress the image noise, so as to obtain a more recognizable image. Finally, the identification and positioning of coal gangue are realized by using AlexNet network.The results show that the recognition accuracy of the system is 95.90%, the accuracy is high, the implementation process is simple and the applicability is good.
作者
何江
张科星
HE Jiang;ZHANG Kexing(Taiyuan University,Taiyuan 030006,China)
出处
《煤炭技术》
CAS
北大核心
2022年第3期205-208,共4页
Coal Technology
基金
山西省高等学校教学改革创新项目(J202177)。