4.1 长距离(相机)轨迹定性实验结果 Qualitative Results on Large Trajectories

We tested the algorithm on several long and challenging trajectories, which include many camera rotations, large scale changes and major loop closures.

我们是在一些长距离,大难度的场景中跟踪(相机运动)轨迹,测试LSD-SLAM算法的,这些(相机运动)轨迹包括了许多相机旋转运动,大的尺度变化以及大闭环(等情况)。

改动_Labby: 我们控制摄像头进行一些长距离、高难度的轨迹运动来检测算法的性能,包括进行多次旋转、大尺度变化以及大闭环运动等。

Figure 7 shows a roughly 500 m long trajectory which takes 6 minutes just before and after the large loop closure is found.

示图(7)显示的是约500米的长距离(相机运动)轨迹,持续时间6分钟,在大回环闭合前后的截图。

Figure 8 shows a challenging trajectory with large variations in scene depth, which also includes a loop closure.

示图(8) 显示的是场景深度变化大的相机运动轨迹,其中还包含闭环情况。

image copy right belongs to engel14eccv paper, 图像摘录自 engel14eccv论文

Fig. 8: Accumulated pointcloud of a trajectory with large scale variation, including views with an average inverse depth of less than 20 cm to more than 10 m.

示图8:(相机运动)轨迹叠加的(半稠密)点云地图,其场景深度的尺度变化大,平均逆深度可以小到20厘米,大到可以超过10米。

After the loop closure (top-right), the geometry is consistently aligned, while before (top-left) parts of the scene existed twice, at different scales.

顶部示图左半部分是闭环前的场景,(两个红色圆圈表示)同样的局部场景在地图中出现两次(即:出现重影),且尺度大小不同。右半部分是大回环闭合之后的场景,(消除了之前的重影现象),使场景具有全局一致性。

改动_Labby: 左半部分——>左上角;右半部分——>右上角。

The bottom row shows different close-ups of the scene.

底部示图显示的是不同尺度下的局部场景特写。(从左到右场景拉伸,使场景尺度变大,译者额外添加备注说明)

The proposed scale-aware formulation allows to accurately estimate both fine details and large-scale geometry – this flexibility is one of the major benefits of a monocular approach.

我们提出的尺度感知公式(主要指 sim(3)\mathfrak{sim}(3) 直接法配准,译者额外添加备注),能够精确地估计场景细节和具有大尺度的场景几何特征—这种灵活性是单目SLAM方法的主要优势之一。

全篇仅提供学习,请勿用于商业用途,翻译版权【泡泡机器人】 all right reserved,powered by Gitbook修订时间: 2017-04-11

results matching ""

    No results matching ""