Large-scale semantic mapping is crucial for outdoor autonomous agents to perform high-level tasks such as planning and navigation. In this paper, we propose a novel method for large-scale 3D semantic reconstruction through implicit representations from posed LiDAR measurements alone. We first leverage an octree-based hierarchical structure to store implicit features, then these implicit features are decoded to signed distance value and semantic information through shallow Multilayer Perceptrons (MLPs). We leverage radial window self-attention networks to predict the semantic labels of point clouds. We then jointly optimize the feature embedding and MLP parameters with a self-supervision paradigm for point-cloud geometry and a pseudo-supervision paradigm for semantic and panoptic labels. Subsequently, geometric structures and object categories for novel points in the unseen area are regressed, and the marching cubes method is exploited to subdivide and visualize scenes in the inferring stage. Experiments on two real-world datasets, SemanticKITTI and SemanticPOSS, demonstrate the superior segmentation efficiency and mapping effectiveness of our framework compared to current state-of-the-art 3D semantic LiDAR mapping methods.

This paper proposes a novel framework for large-scale scene reconstruction based on 3D Gaussian splatting (3DGS) and aims to address the rendering deficiency and scalability challenges faced by existing embodied AI tasks. For tackling the scalability issue, we split the large scene into multiple cells, and the candidate point-cloud and camera views of each cell are correlated through a visibility-based camera selection and a progressive point-cloud extension. To reinforce the rendering quality, three highlighted improvements are made in comparison with vanilla 3DGS, which are a strategy of the ray-Gaussian intersection and the novel Gaussians density control for learning efficiency, an appearance decoupling module based on ConvKAN network to solve uneven lighting conditions in large-scale scenes, and a refined final loss with the color loss, the depth distortion loss, and the normal consistency loss. Finally, the seamless stitching procedure is executed to merge the individual Gaussian radiance field for novel view synthesis across different cells. Evaluation of Mill19, Urban3D, and MatrixCity datasets shows that our method consistently generates more high-fidelity rendering results than state-of-the-art methods of large-scale scene reconstruction. We further validate the generalizability of the proposed approach by rendering on self-collected video clips recorded by a commercial drone.