摘要
针对露天矿山炮孔识别技术进行了深入研究,提出并比较了2种炮孔智能识别方法:基于三维点云的炮孔识别方法和基于目标检测的炮孔识别方法。通过选用不同视觉感知设备,在多个露天矿山环境下进行了试验验证和性能分析。研究发现,2种方法均能有效识别炮孔,具体来说,基于三维点云的炮孔识别方法识别精确率为90%,基于目标检测的炮孔识别方法识别精确率为97.91%。通过对2种方法的硬件设备、数据处理流程和应用潜力进行详细比较,结果表明,结合人工智能技术,炮孔识别技术在智能化炸药现场混装车等采矿装备中具有广阔的应用前景,对推动采矿技术进步,实现无人、高效、安全的矿山开采模式具有重要理论与实际意义。
An in-depth study on blast hole recognition technology in open-pit mines is conducted,proposing and comparing 2 intelligent blast hole recognition methods:a 3D point cloud-based method and a target detection-based method.Experiment verification and performance analyses were carried out in various open-pit mining environments using different visual perception devices.The study reveals that both methods effectively recognize blast holes,with the 3D point cloud-based method achieving a recognition accuracy of 90%,and the target detection-based method achieving an accuracy of 97.91%.Detailed comparisons of hardware devices,data processing workflows,and application potential between the 2 methods indicate that when integrated with artificial intelligence technologies,blast hole recognition technology holds significant promise for applications in intelligent on-site bulk charging trucks and other mining equipment.This advancement plays a vital theoretical and practical role in promoting mining technology,enabling unmanned,efficient,and safe mining operations.
作者
胥维
迟洪鹏
战凯
Xu Wei;Chi Hongpeng;Zhan Kai(BGRIMM Technology Group;BGRIMM Explosives&Blasting Technology Co.,Ltd.)
出处
《黄金》
2025年第1期6-10,共5页
Gold
基金
国家重点研发计划项目(2023YFC2907402)。
作者简介
胥维(1996—),男,工程师,硕士,研究方向为计算机视觉、深度学习、智能爆破,E-mail:xuwei@bgrimm.com;通信作者:迟洪鹏(1983-),男,正高级工程师,博士,从事智能化工业炸药现场混装技术与装备、机器人、智能工厂等方面的技术研究工作,E-mail:hongpengchi@163.com。
