Complementary Method for Discontinuous Drone Odometry and Improved Mapping Accuracy Using 1-Channel LiDAR

This study proposes an offline processing method to address discontinuous drone odometry and improve mapping accuracy in terrestrial mapping applications using single-channel LiDAR. Odometry discontinuities frequently occur during drone-based data logging due to hardware constraints such as on- board computer overheating, communication instability, and processing delays. To overcome these limitations, this study extracts ORB-SLAM3-based visual odometry from logged camera images and uses a conditional complement strategy to selectively replace drone odometry only in discontinuous sections, ensuring continuity while maintaining accuracy. Furthermore, we construct individual maps for each flight path and apply ICP-based map-to-map matching to compensate for z-value errors due to the vertical accuracy limitations of RTK. Experimental results demonstrate that the proposed approach significantly improves continuity while maintaining accuracy comparable to drone odometry. Quantitative evaluation results demonstrate that ICP matching significantly reduces z-value errors across all experimental regions. This offline processing method produces maps with higher density than conventional methods while maintaining accuracy, effectively providing reliable topographic information despite the limitations of single-channel LiDAR systems. The results of this study will contribute to the development of practical high-precision mapping systems applicable to various fields, including agriculture and construction, in resource-constrained environments.