As an optical sensor, LiDAR cannot effectively detect transparent obstacles in the environment, such as glass parapets and glass doors in buildings. In this article, a glass detection and point cloud reconstruction method relying only on LiDAR is proposed, which can improve the detection success rate for glass and achieve the reconstruction of glass. The algorithm achieves the detection of glass point clouds by reflection intensity and local structure feature of point clouds. Experiments prove that the algorithm can well identify the glass position and reconstruct the point cloud of glass. It lays the foundation for mobile robots to carry out simultaneous localization and mapping (SLAM) and path planning tasks smoothly in the indoor environment with glass.