为了实现煤矿井下综采工作面的无人化开采，代替工作人员监测危险的开采环境，设计了一款基于激光SLAM的矿用巡检装置控制系统。针对低功耗、并行处理快速等需求，选择NVIDIA Jetson AGX作为主控模块，并采用WiFi和LoRa双通信方式以满足不同的数据传输需求；驱动系统核心采用STM32F103C8T6微控制器，通过接收主控模块指令驱动直流无刷电机；算法采用非极大值抑制精细化特征提取，以实现局部最大搜索，从而更精准地定位和识别目标。在室内实验室场景与煤矿井下综采工作面2种不同的实验环境中对巡检装置控制系统的建图清晰度、精确定位和标靶识别效果进行了样机测试。试验结果表明，设计的建图效果在特征提取等算法方面表现出色，具有较高的定位精度，满足了实际应用需求。

In order to realize the unmanned mining of the fully mechanized mining face in the coal mine, instead of the staff to explore and monitor the dangerous mining environment, we designed a control system for a mining inspection device based on laser SLAM. To meet the requirements of low power consumption and rapid parallel processing, NVIDIA Jetson AGX was chosen as the main control module, and dual communication method of WiFi and LoRa was adopted to meet different data transmission needs. The core of the driving system employs the STM32F103C8T6 microcontroller to drive the brushless DC motor by receiving instructions from the main control module. The algorithm uses non-maximum suppression to extract fine feature to achieve local maximum search, so as to locate and identify the target more accurately. Prototype tests of the inspection device control system were conducted in two different experimental environments: an indoor laboratory setting and an underground coal mine comprehensive mining working face. The tests aimed to assess mapping clarity, precise positioning, and target recognition effectiveness. The experimental results show that the design has excellent performance in feature extraction and other algorithms, and has high positioning accuracy, which meets the needs of practical applications.