The benchmark of LiDAR odometry algorithms utilised for a low-cost mobile mapping system

Abstract. In recent years, the development of Mobile Mapping Systems (MMS) has seen a noticeable increase, particularly those based on low-cost sensors for acquiring various geospatial data. This has contributed to the growth in the number of different applications, and mobile mapping is now widely used, for example, in autonomous vehicles, photogrammetric Unmanned Aerial Vehicle (UAV) missions, and various navigation applications and the MMS is also considered a method of mobile space inventory. Another significant application of MMS is the continuous development of building interior mapping technology to analyse the construction process and create applications that support the movement of people, especially for emergency services or people with disabilities. A benchmark of current SOTA LiDAR odometry algorithms was proposed. It is available at MapsHD/HDMapping [https://github.com/MapsHD/HDMapping]. A benchmark of 20 algorithms: CT-ICP, DLIO, DLO, FASTER-LIO, FAST-LIO, GenZ-ICP, GLIM, i2ekf-lo, ig-lio, KISS-ICP, LeGO-LOAM, LiDAR_IMU_Init, LIO-EKF, LIO-SAM, LOAM, MAD-ICP, POINT-LIO, RESPLE, SLICT and VOXELMAP was conducted in the laboratory and challenging Bunker DVI Dataset.