课题基金 / 基金详情

矿用防爆相机折射成像建模与多激光束掘进设备视觉定位方法

批准号:
52104166
项目类别:
青年科学基金项目(C类)
资助金额:
30.0 万元
负责人:
杨文娟
依托单位:
学科分类:
智能矿山
结题年份:
2024
批准年份:
2021
项目状态:
已结题
项目参与者:
杨文娟

项目摘要

结项摘要

杨文娟的其他基金

相似基金

相关文献

中文摘要
针对防爆玻璃折射图像畸变导致井下视觉定位系统精度和稳定性不足问题,提出一种考虑玻璃折射成像机理及建模理论,利用多激光束构建井下点-线定位特征,获得非透视多激光束精确定位的新方法。依托已有平台和研究基础,研究具有多层防爆玻璃折射影响的矿用相机非中心折射成像机理,构建几何驱动的多轴矿用相机共面约束成像模型;建立矿用相机折射校准近似本质矩阵方程,以重投影量化误差最小为目标,实现矿用相机参数非线性估计与优化;以三个平行激光指向仪的出射激光点和激光线构建测量特征,建立非透视三点三线定位模型,实现掘进设备长距离稳定精确定位;建立基于距离范围和测量精度的非线性多参数优化模型,优化井下视觉定位系统方案。在所研制多激光束标靶掘进设备位姿测量平台上进行性能实验研究,验证并利用多因素耦合下矿用相机非线性成像行为,获得煤矿井下设备视觉测量系统设计准则,为高精度和鲁棒性强的井下定位系统设计奠定理论和技术基础。
英文摘要
Aiming at the problems of the accuracy and stability of the visual positioning system that result from the image distortion caused by the refraction of explosion-proof glass, a modeling theory considering the mechanism of glass refraction imaging is proposed, and multiple laser beams have been used to construct the point-line positioning features in coal mine, so as to obtain the precise positioning of non-perspective multiple laser beams. Based on the existing platform and research foundation, the non-central refraction imaging mechanism of underground camera with the refraction effect of multi-layer explosion-proof glass is studied, and a geometrically-driven coplanar constrained imaging model of multi-axis underground camera is constructed. The similar essential matrix equation of refraction calibration for underground camera is established, and the nonlinear estimation and optimization of underground camera parameters are realized with the goal of minimizing re-projection quantization error. The measurement features are constructed by the outgoing laser points and laser lines of three parallel laser pointing instruments, and the non-perspective three-point and three-line positioning model is established to realize the long distance stable and accurate positioning of tunneling equipment. A nonlinear multi-parameter optimization model based on the range and measurement accuracy is established to optimize the vision positioning system scheme. The performance experiments have been carried out on the position and posture measurement platform of the tunneling equipment with multiple laser beams target, and the nonlinear imaging behavior of the underground camera under the coupling of multiple factors has been verified. The design criteria of the vision measurement system for the underground equipment in coal mine are obtained, which have laid the theoretical and technical foundation for the design of the positioning system with high precision and strong robustness.
本项目紧密结合煤矿智能化领域的研究热点和井下巷道掘进装备精确位姿测量困难的行业需求,针对防爆玻璃折射图像畸变导致井下视觉定位系统精度和稳定性不足问题,提出一种考虑玻璃折射的矿用防爆相机非线性成像机理、建模与标定方法,为提升掘进设备定位精度和稳定性提供有力支撑。项目研究了矿用相机成像机制驱动的多激光束井下精确定位理论和方法,开展了具有多层防爆玻璃折射影响的矿用相机非中心折射成像机理、几何驱动的多轴矿用相机共面约束成像模型及矿用相机折射校准参数非线性估计与优化、多激光束点-线特征亚像素级精确提取及掘进设备三点三线定位等方面的基础研究,搭建了掘进设备定位系统性能测试实验平台。测试结果表明:多轴矿用防爆相机折射成像模型可以有效表征煤矿井下视觉测量系统的非中心折射畸变,提出的校正理论和方法有效解决了矿用相机玻璃折射导致的图像退化问题;提出的井下视觉系统非单视点折射畸变校准模型,实现了矿用相机模型估计参数的优化与修正;提出了基于多尺度注意力机制的语义分割网络RCEAU-Net模型,实现了多激光束图像特征亚像素级精确提取;提出的基于多激光束点-线特征的掘进设备三点三线(3P3L)定位模型,通过建立测量模型的最小二乘法损失函数,结合最小化重投影误差实现了非迭代全局最优解估计。开展了不同工况下掘进设备定位系统精度与稳定性测试,验证了所提出的掘进设备视觉定位理论及方法。上述理论及方法已经用于陕西彬长集团孟村煤矿、新疆宝英煤矿掘进机定位系统,井下工业测试达到了施工要求。项目研究对煤矿井下高精度视觉定位系统设计具有重要参考价值,为进一步煤矿井下智能掘进定位与导航系统应用推广奠定了坚实的理论基础。
矿用分光式事件相机异步协同成像机理及掘进设备融合定位方法
  • 批准号:
    52574204
  • 项目类别:
    面上项目
  • 资助金额:
    50.00万元
  • 批准年份:
    2025
  • 负责人:
    杨文娟
  • 依托单位:
国内基金
海外基金