Estimated trajectory data from our work

This dataset was used to support our work and provided to the review for reference. The TUM RGB-D dataset comprises image sequences from dynamic indoor settings, containing various line densities and scene dynamics conducive for SLAM analysis. We selected four walking and two sitting sequences to demonstrate high and low dynamics, respectively. The walking sequences (w/half, w/rpy, w/static, w/xyz) capture two individuals in motion, including chair movement, while the sitting sequences (s/half, s/xyz) document two individuals in conversation with minimal movements. The appended terms `half', `rpy', `static', and `xyz' specify different camera movements. The KITTI dataset consists of images in natural and urban outdoor settings. The KITTI dataset is a collection of images captured in varied natural and urban outdoor environments. Our research specifically zeroes in on sequences 00 to 10, all of which supply us with verified ground truth data. The sequences we've chosen for analysis are characterized by their unique mix of traffic density and the prevalence of man-made structures. By conducting a comprehensive examination of these sequences, we are able to gauge the effectiveness and robustness of our DPL-SLAM system across an array of different driving environments.

本数据集用于支撑本研究工作，并提交至评审环节以供参考。TUM RGB-D 数据集（TUM RGB-D）包含来自动态室内场景的图像序列，涵盖多种线条密度与场景动态变化特征，适用于同步定位与地图构建（SLAM）分析。我们选取了4段行走序列与2段静坐序列，分别用于展示高动态与低动态场景。行走序列（w/half、w/rpy、w/static、w/xyz）记录了两名运动中的个体，还包含椅子移动的场景；静坐序列（s/half、s/xyz）则记录了两名以极小幅度动作进行交谈的个体。后缀术语`half'、`rpy'、`static'与`xyz'用于区分不同的相机运动模式。KITTI 数据集（KITTI）由多种自然与城市户外环境下采集的图像组成。本研究重点聚焦于序列00至10，所有这些序列均提供了经过验证的真实标注（ground truth）数据。我们选定用于分析的这些序列，其特征在于兼具多样的交通密度与丰富的人工建筑场景。通过对这些序列开展全面测试，我们能够评估DPL-SLAM 系统（DPL-SLAM）在多种不同驾驶环境下的有效性与鲁棒性。