点击下方 卡片 ，关注" 自动驾驶之心 "公众号 戳我-> 领取 自动驾驶近30个 方向 学习 路线 >>直播和内容获取转到 → 自动驾驶之心知识星球 点击按钮预约直播 近年来，3DGS在自动驾驶领域的城市场景重建中展现出巨大潜力。然而，当前的城市场景重建方法通常依赖于多模态传 感器输入，例如激光雷达和图像。尽管激光雷达点云提供的几何先验在很大程度上可以缓解重建中的不适定性，但在实 践中获取如此精确的激光雷达数据仍然具有挑战性： i）需要对激光雷达与其他传感器之间进行精确的时空标定，因为它们可能并非同时捕获数据；ii）当激光雷达和相机安 装在不同位置时，空间未对准会导致重投影误差。为了避免获取精确激光雷达深度的困难，本文提出了 D²GS， 一种无需 激光雷达的城市场景重建框架。 在这项工作中，获得了与激光雷达效果相当，但更密集、更精确的几何先验。本次自动 驾驶之心为大家邀请到D²GS作者 博世创新软件中心，三维重建算法专家张友健 为大家详细分析本篇工作。 传统的自驾场景重建方案依赖激光雷 达作输入,往往会遇到标定误差和深 度投影误差等问题。对此，博世合成 数据团队提出,通过多视图深度初始 化高斯点云，并在训练 ...

Core Viewpoint - The article discusses the introduction of DriveSplat, a new method for 3D reconstruction of driving scenes that significantly enhances the accuracy of both static and dynamic elements, achieving state-of-the-art performance in novel view synthesis tasks on two autonomous driving datasets [2][41]. Group 1: Background and Motivation - Realistic closed-loop simulation of driving scenes has become a major research focus in both academia and industry, addressing factors such as fast-moving vehicles and dynamic pedestrians [2][5]. - Traditional methods have struggled with motion blur and geometric accuracy in dynamic driving scenes, leading to the development of DriveSplat, which utilizes a decoupled approach for high-quality scene reconstruction [2][6][7]. Group 2: Methodology - DriveSplat employs a neural Gaussian representation with a decoupled strategy for dynamic and static elements, enhancing the representation of close-range details through a partitioned voxel initialization scheme [2][8][14]. - The framework incorporates deformable neural Gaussians to model non-rigid dynamic participants, with parameters adjusted over time using a learnable deformation network [2][8][21]. - The method leverages depth and normal priors from pre-trained models to improve geometric accuracy during the reconstruction process [2][23][41]. Group 3: Performance Evaluation - DriveSplat was evaluated on the Waymo and KITTI datasets, demonstrating superior performance in both scene reconstruction and novel view synthesis compared to existing methods [28][31]. - In the Waymo dataset, DriveSplat achieved a PSNR of 36.08, surpassing all baseline models, while also showing improvements in SSIM and LPIPS metrics [28][29]. - The method also outperformed competitors in the KITTI dataset, particularly in maintaining background detail and accurately rendering dynamic vehicles [31][32]. Group 4: Ablation Studies - Ablation studies indicated that the combination of SfM and LiDAR for point cloud initialization yielded the best rendering results, highlighting the importance of effective initialization methods [33][34]. - The background partition optimization module was shown to enhance performance, confirming its necessity in the reconstruction process [36]. - The introduction of a deformable module significantly improved the rendering quality of non-rigid participants, demonstrating the effectiveness of the dynamic optimization approach [39][40].

Core Viewpoint - The article discusses the development of Bezier curve Gaussian splatting (BezierGS) by Fudan University, which addresses the challenges of dynamic target reconstruction in autonomous driving scenarios, improving the accuracy and efficiency of scene element separation and reconstruction [1][2]. Group 1 - BezierGS utilizes learnable Bezier curves to represent the motion trajectories of dynamic targets, leveraging temporal information to calibrate pose errors [1]. - The method introduces additional supervision for dynamic target rendering and consistency constraints between curves, leading to improved reconstruction outcomes [1]. - Experiments on the Waymo Open Dataset and nuPlan benchmark demonstrate that BezierGS outperforms state-of-the-art alternatives in both dynamic and static scene target reconstruction [1]. Group 2 - The article highlights the potential to build a high-quality street scene world for training and exploring self-driving models, which can reduce data collection costs [2]. - It emphasizes the reduction of reliance on the accuracy of bounding box annotations, which are often imprecise in current industry and open-source datasets [2]. - The work represents a step towards exploring a true self-driving world model, although it currently only achieves trajectory interpolation and not extrapolation [2].