Abstract: Objective　With the widespread use of natural gas in daily life and urban infrastructure, the safe and reliable operation of urban gas pipeline networks is essential for protecting public safety and property. Laser-based detection technology has become an important approach for routine safety monitoring and maintenance of gas pipeline systems. Vehicle-mounted methane remote sensing equipment based on tunable diode laser absorption spectroscopy (TDLAS) offers rapid response, high detection sensitivity, and non-contact, long-range detection, making it suitable for mobile gas inspection. However, urban gas pipeline networks typically cover extensive areas with uneven spatial distributions and contain large volumes of spatial data. Moreover, vehicle-mounted inspection requires pipeline network data to be processed and visualized in real-time to accommodate normal driving speeds. Under such conditions, conventional systems often suffer from delays in pipeline rendering and positioning, which seriously hinder the rapid identification and localization of gas leakage points. Therefore, achieving real-time and efficient visualization of pipeline network data remains a key technical challenge for vehiclemounted gas inspection. To address this challenge, this study integrates vehicle-mounted methane remote sensing equipment with data simplification and spatial indexing techniques to develop an intelligent vehicle-mounted gas inspection system. The proposed system enables efficient real-time visualization of pipeline networks during mobile inspection and accurate localization of gas leakage points under vehicle-mounted monitoring conditions. This study advances the monitoring technology of vehicle-mounted gas pipeline and provides technical support for urban gas safety management and public safety. Methods　To address the challenge of rapid visualization of massive pipeline datasets in conventional vehicle-mounted inspection systems, this study optimizes the system from two key aspects: data query efficiency and graphic rendering efficiency. On the one hand, grid partitioning is employed to construct the minimum bounding rectangle (MBR) of regional pipeline data, and then an R-tree spatial index is constructed based on the MBRs to effectively improve the query efficiency of target pipeline area. On the other hand, a multi-way tree structure is constructed to describe the topological relationship of pipeline segments within each MBR. After acquiring the ordered set of pipeline segments, the sliding-window algorithm is used for adaptive data simplification. This method can effectively reduce redundant data while maintaining the overall morphological characteristics of the original pipeline network, thereby realizing fast and high-quality visual rendering of natural gas pipeline networks on the intelligent gas inspection and operation platform. Results and Discussion　All experiments were conducted in Windows 10 environment with .NET Framework 4.5. Actual gas pipeline network data of Langfang City were used to evaluate the proposed method. An R-tree index was constructed for the pipeline data, and the effects of different threshold values on pipeline data simplification were systematically evaluated. Multiple evaluation indicators, including geometric features, positional accuracy, and the number of simplified line segments, were considered to determine the optimal distance threshold. A distance threshold value of 6 m was finally selected because it provided an effective balance between the simplification degree and the maintaining of integrity of the pipeline network geometry. It is shown that when a 12 MB pipeline network dataset was tested, the proposed system achieved an average map response time of 76 ms. This result demonstrates that the proposed approach can substantially improve the efficiency of pipeline network visualization while maintaining the essential geometric characteristics of pipeline networks, fully meeting the practical application requirements of vehicle-mounted gas inspection. Conclusions　Focusing on vehicle-mounted laser remote sensing for natural gas inspection, this study investigated the rapid visualization of pipeline networks under mobile inspection conditions.A depth-first search (DFS) matching strategy was adopted to establish accurate topological relationships among pipeline data. This strategy recursively traverses pipe segment endpoints and matches adjacent geometric features according to coordinate proximity, thereby reducing the reliance on external platforms for data preprocessing and improving the independence of the inspection system. In contrast to conventional approaches that usually require pre-generating topological connections using external GIS platforms, the proposed method directly extracts topological connections from original pipeline geometric data. By combining an R-tree spatial index with a sliding-window data simplification algorithm, the proposed method improves both pipeline data query efficiency and graphic rendering efficiency, thus substantially reducing visualization delay caused by large-scale spatial datasets and meeting the real-time requirements of vehicle-mounted natural gas detection. The proposed algorithm also demonstrates good software portability, providing a practical and reliable approach for the efficient visualization of urban gas pipeline networks and supporting the broader application of laser remote sensing technology in gas pipeline inspection.

Key words: navigation detection, R-tree, sliding window algorithm, deep search matching, topological relation