A 2.5D SLAM Algorithm Based on the Fusion of 2D LiDAR and RGB-D Camera

In multi-sensor fusion SLAM systems, the integration of 2D LiDAR and RGB-D camera point clouds often yields promising results. This paper employs progressive morphological filtering techniques to process the point cloud data generated by the RGB-D camera, effectively filtering out ground points and preventing the misclassification of ground surfaces as obstacles during the fusion of camera and LiDAR point clouds, thereby improving the accuracy of environmental perception. Additionally, a global point cloud processing strategy is adopted to retain key and stable camera points that remain consistent across different perspectives, enhancing the system's robustness and localization accuracy in dynamic environments. By fusing data from these two types of sensors, the system achieves more reliable environmental perception and map construction.