Derkai
Derkai
You can learn Rviz by simply clicking on "Add" -> "topic" and you will get the answer you want.
This is probably normal, and I've found that this only happens at the end of a batch
More precisely, in Carla0.9.14, this should be changed to `point_cloud.append([location.point.y, -location.point.x,-location.point.z]) ` to ensure that the visualization is correct at the same time
Hello, it may be the cause of slow computer performance, you need to be patient or reduce the rendering quality to achieve your work
是的，我已经尝试你的仓库了，它运行成功，但是精度非常差
> Hey Teddy, > > I've added LighterGlue ONNX export in my fork here: [stschake@acecc15](https://github.com/stschake/accelerated_features/commit/acecc1561a60fc30904f0a79e0def825fc7224d1) Thanks！
我目前正在尝试复现EdgePoint+Lightglue
> > 您好，感谢你的问题。fast-livo2的点云上色策略与fast-livo1保持一致，即直接将每帧的原始点 (raw points) 通过优化后的位姿投影到对应图像上进行赋色，然后发布到rviz中进行可视化。相比于**后处理的多帧点云上色**或**使用贝叶斯更新来平滑地图颜色**，这种方式能够更加直观地暴露位姿估计的问题。只要pose存在细微的误差，彩色点云就会出现模糊，从而更快定位问题。 > > 请问发布到rviz中赋色之后的点云，如何存储到本地呢？ 在对应启动的 xxxx.yaml 文件中，打开： pcd_save: pcd_save_en: true 就能保存在 Log/PCD 下面了，如果没有这个文件夹需要自己创建一下，在程序结束后，一共会自动生成两个点云文件，都是带RGB的点云，只不过一个是降采样的稀疏点云，一个是原始点云。