首页|期刊导航|机械与电子|炮孔视觉定位与机械臂装填系统设计

炮孔视觉定位与机械臂装填系统设计OA

Design of a Visual Positioning and Robotic Arm Charging System for Blast Holes

中文摘要英文摘要

针对隧道掌子面弱光、非均匀照明、岩壁纹理复杂以及孔缘破损条件下孔位识别与机械臂对准不稳定的问题,设计了一种面向隧道炮孔的视觉定位与机械臂自动装填系统.系统以工业相机采集的掌子面图像为输入,首先通过低照度增强、全孔候选提取和空间分布约束获得炮孔整体分布;随后在中心作业孔感兴趣区域(ROI)内构建候选孔综合评分模型和四孔几何代价函数,并结合椭圆拟合完成孔心精修;最后利用单应性映射将孔心坐标转换为机械臂作业平面坐标,采用安全趋近、孔口对准、轴向直线插补和复位的分阶段控制流程完成模拟药卷自动装填.实验结果表明,在 20 张掌子面图像、1 140 个标注炮孔条件下,所提方法全孔识别 F1 值达到 0.992 5,中心四孔平均定位误差约为 1.74 mm;在 60 次模拟装填实验中,系统综合成功率为 95.0%,单孔平均耗时约 8.2 s.结果表明,该系统能够为复杂弱光环境下的炮孔视觉定位与机械臂自动装填提供可行的机电一体化方案.

To address the challenges of hole recognition and unstable robotic arm alignment under con-ditions of low light,non-uniform illumination,complex rock wall textures,and damaged hole edges in tun-nel faces,this paper designs a visual positioning and automatic charging system for tunnel blast holes.The system takes tunnel-face images captured by an industrial camera as input.First,the low-light enhance-ment,the full-hole candidate extraction,and the spatial distribution constraints are applied to obtain the overall distribution of blast holes.Subsequently,within the region of interest(ROI)of the central operat-ing holes,a comprehensive scoring model for candidate holes and a geometric cost function for the four-hole pattern are constructed,and combined with ellipse fitting to refine the hole centers.Finally,the ho-mography mapping transforms the hole-center coordinates into the working plane coordinates of the ro-botic arm.A staged control strategy,comprising safe approach,hole alignment,axial linear interpolation,and resetting,is adopted to accomplish automatic charging of simulated explosive cartridges.Experimental results show that,on 20 tunnel-face images with 1 140 annotated blast holes,the proposed method a-chieves an F1-score of 0.992 5 for full-hole recognition,with an average localization error of approxi-mately 1.74 mm for the central four holes.In 60 simulated charging trials,the overall system success rate reaches 95.0%,and the average time per hole is about 8.2 seconds.These results indicate that the system provides a feasible mechatronic solution for visual positioning of blast holes and automatic robotic charging in complex low-light conditions.

赵亚龙;张周强;常旭冰;张博

西安工程大学机电工程学院,陕西 西安 710048西安工程大学机电工程学院,陕西 西安 710048西安工程大学机电工程学院,陕西 西安 710048西安工程大学机电工程学院,陕西 西安 710048

信息技术与安全科学

机器视觉机械臂炮孔定位自动装填

machine visionrobotic armblast-hole localizationautomatic filling

《机械与电子》 2026 (9)

66-73,8

陕西省教育厅科学研究计划高校工程研究中心项目(24JR072)功能性纺织材料及制品教育部重点实验室资助项目(2024FTMP016)陕西省自然科学基础研究计划项目(2026JC-YBMS-0033)

评论