基于UNet的干式磁选分选颗粒轨迹在线检测识别方法
Method for online detection and recognition of particle trajectories in dry magnetic separation based on UNet
针对干式磁选分选过程中颗粒形状、大小和颜色等特征多样性所带来的颗粒轨迹检测和识别难点，提出了一种基于UNet的颗粒轨迹在线检测识别方法。该方法采用深度学习技术对干式磁选分选过程的图像序列进行处理，快速识别颗粒轨迹，并提取分析颗粒特征。实验结果表明，该方法的识别准确率高达93.5%上，有效提高了干式磁选分选过程的效率和精度，降低了人工操作的成本和误差率。该方法还为干式磁选分选过程的精细化管理提供了技术支持。其现实意义在于提高干式磁选分选过程的自动化和智能化水平，为矿山企业的高效、稳定和可持续发展提供支持。本文的研究方法和结论可以为其他颗粒轨迹检测和识别领域提供借鉴和参考。
The article proposes a UNet-based online detection and recognition method for particle trajectories to address the challenges of detecting and identifying particle trajectories in dry magnetic separation processes due to the diversity of particle characteristics, such as shape, size, and color. The method processes image sequences in dry magnetic separation processes using deep learning technology to quickly identify particle trajectories and extract and analyze particle features. The experimental results show that the method achieves a high recognition accuracy of over 93.5% effectively improving the efficiency and accuracy of dry magnetic separation processes and reducing the cost and error rate of manual operations. The method also provides technical support for the refined management of dry magnetic separation processes. Its practical significance lies in improving the automation and intelligence level of dry magnetic separation processes and providing support for the efficient, stable, and sustainable development of mining enterprises. The research methods and conclusions can serve as a reference for other fields of particle trajectory detection and recognition.
黄勇;邹立超;夏星;肖盛旺;
长沙矿冶研究院 湖南 长沙 410012;长沙矿冶研究院 湖南 长沙 410012;长沙矿冶研究院 湖南 长沙 410012;长沙矿冶研究院 湖南 长沙 410012;
Dry magnetic separation Particle trajectory detection UNet Deep learning
1-2 / 2
评论