为解决隧道爆破质量控制困难的问题，探究新型爆破参数优化方法，提出一种3D激光扫描技术与BP神经网络相结合的实时优化方法，对隧道爆破超欠挖感知及其后续钻爆参数进行实时修正，构建现场扫描方法、数据处理、点云提取与超欠挖图像等工作程序，建立BP神经网络模型、算法并确定模型参数。通过试验验证表明： 1）将3D激光扫描技术应用于隧道爆破超欠挖质量感知中切实可行，与人工测量结果吻合较好，具有快速、实时和准确的优点； 2）采用3D激光扫描方法能够对隧道爆破后的断面轮廓线进行精确评测，在获取爆破超欠挖的精细数据后，通过样本学习完成模型训练能推理出较为合理的爆破参数，改进后的优化方案使平均超挖降低54.8%。

During tunnel blasting construction, the blasting quality is difficult to control. Therefore, a real-time blasting parameter optimization method based on three-dimensional(3D) laser scanning technology and back propagation(BP) neural network is proposed. This method percepts the over-and under-excavation of tunnel blasting and corrects subsequent drilling and blasting parameters in real time. The working procedures of field scanning method, data processing, point cloud extraction, and over- and under-excavation images are constructed. Finally, the BP neural network model, algorithm, and model parameters are established. The verification test results show that: (1) It is feasible to apply 3D laser scanning technology to the over- and under-excavation quality perception of tunnel blasting, which is in good agreement with the manual measurement results, and has the advantages of fast, real-time, and accurate. (2) The 3D laser scanning method can accurately evaluate the profile line of the tunnel after blasting. After obtaining precise data on blasting over- and under-excavation, the model training can be completed through sample learning to infer reasonable blasting parameters. The improved optimization scheme reduces the average over excavation by 54.8%.