煤岩界面识别技术是煤矿智能化开采的关键技术之一。基于高频雷达波探测技术可实现煤岩界面的随采高精度探测，但仍存在矿井超大采高(≥6 m)片帮垮落带来设备的安全风险及采高突变(采高≤2 m)时空间限制设备通过的问题。在前期工作基础上提出了一种矿井煤岩界面节点式雷达快速动态探测系统并进行了煤岩界面探测实验研究，主要内容包括：① 阐述矿井节点式雷达观测系统原理，根据矿井工作面实际环境设计煤岩界面识别观测系统方案及雷达传感单元安装方式；② 研究并提出节点式采集控制系统和信息交互传输设计方案，实现数据动态采集控制及存储；③ 针对节点式采集方式及煤岩界面雷达反射回波特征，研究提出了节点探测数据增强处理方法、煤岩界面识别算法，可有效的实现煤岩界面智能识别与追踪、煤层厚度及空间坐标解算。为验证该方法的可行性，采用多个中心频率为1.5 GHz的探地雷达传感单元进行物理模型验证实验，并对节点式数据采集和连续数据采集结果进行了对比分析，实验结果表明：节点式采集方法与连续采集方法均可有效识别出煤岩界面，与连续采集方法相比，本文提出的节点式探测方法可实现数据的快速动态重复性采集，单次采集时长控制在10 s以内，煤层厚度探测结果平均误差为1.07 cm，最大误差为1.47 cm，平均误差百分比为7.64%。本方法为矿井智能化开采中煤岩界面的动态高精度探测提供技术支撑。

Coal-rock interface recognition technology is one of the key technologies for intelligent mining in coal mines. Based on high-frequency radar wave detection technology, high-precision detection of coal-rock interfaces can be achieved with mining, but there are still safety risks for equipment caused by rib spalling and roof caving in ultra-high mining heights (≥6 m) in mines, as well as spatial restrictions on equipment passing through during sudden changes in mining height (mining height ≤2 m). Based on previous work, this paper proposes a rapid dynamic detection system and method for coal-rock interfaces in mines using nodal GPR. The main contents include: ① Explaining the principle of the nodal GPR observation system in mines, designing a coal-rock interface recognition observation system plan and radar antenna sensor installation method based on the actual environment of the mine working face; ② Studying and proposing a nodal acquisition control system and information interaction transmission design plan to achieve dynamic data acquisition, control, and storage; ③ Studying and proposing enhanced processing methods for sensor detection data and coal-rock interface recognition algorithms based on nodal acquisition methods and radar reflection echo characteristics of coal-rock interfaces, which can effectively achieve intelligent recognition and tracking of coal-rock interfaces, as well as calculation of coal seam thickness and spatial coordinates. To verify the feasibility of this method, multiple geological radar antenna sensors with a center frequency of 1.5 GHz were used for physical model verification experiments, and a comparative analysis of nodal data acquisition and continuous data acquisition results was conducted. The experimental results show that both nodal and continuous acquisition methods can effectively identify coal-rock interfaces. Compared with continuous acquisition methods, the nodal detection method proposed in this paper can achieve rapid and dynamic repetitive data acquisition, with a single acquisition time controlled within 10 seconds, an average error of 1.07 cm for coal seam thickness detection results, a maximum error of 1.47 cm, and an average error percentage of 7.64%. This method provides technical support for dynamic high-precision detection of coal-rock interfaces in intelligent mining in mines.