Comments (3)
Hi @hitsjt ,
So the getFrustumCulledVoxels function is simply trying to minimize wasted time. Since we know the focal parameters and location of the camera, we know which voxels are not in its field of view. This means we can "cull" (i.e. trim down) only to the voxels we actually will need to update.
When we deal with poses, we tend to use what I'd call "Computer Vision" coordinates: X is right, Y is down, Z is forward. The PCL class we use for computing the view frustum, pcl::FrustumCulling, seems to use what I'd call "Robotics" coordinates, flipping X and Z and negating Y. This is just a coordinate transform from one arbitrary system to another:
Eigen::Matrix4f trans_robot = trans.matrix ().cast () * cam2robot;
from cpu_tsdf.
@hitsjt sorry! This is why I shouldn't have turned github alerts off :(.
I'm guessing by now your questions are no longer relevant, but for posterity:
- Yes, I transform the voxel center to the current frame and reproject it to get a new observation
- This new "distance" is the new observation of d, where d is "how far, in the Z dimension, is the point from the voxel center?" If the point had been exactly at the voxel center, this would be an observation of 0, or a perfect reading of the surface.
- My split-checking is all something I came up with, to try to make these things fit in memory. I keep big voxels and only iteratively make them smaller as needed. So if a point falls within a voxel, I check to see if I need to split it, then recursively check each split voxel for their own readings.
- The coordinates are a bit arbitrary, just things I happened to notice work well for walking around a room with a Kinect. The camera is always at the origin (0,0,0) for me when I begin, but of course, you input that yourself whenever you call integrateCloud(). In general people record things by standing a bit away and pointing. So assuming it was about 2 meters in front of the camera seemed to work well.
from cpu_tsdf.
Thank you sdmiller!
I still have more questions.
- In function
template <typename PointT, typename NormalT> int cpu_tsdf::TSDFVolumeOctree::updateVoxel ( const cpu_tsdf::OctreeNode::Ptr &voxel, const pcl::PointCloud<PointT> &cloud, const pcl::PointCloud<NormalT> &normals, const Eigen::Affine3f &trans_inv)
,- you transformed voxel center to current frame,and reproject it to get the current observation,am i right?
- Then you do
float d_new = (pt.z - v_g.z);
to get the new distance,why is that? - and then what is the purpose of checking split in the following code?
- Question about the coordinates
If i am not wrong ,you put the whole volume center at (2,2,2),but where is the cam?
Thank you!
from cpu_tsdf.
Related Issues (20)
- Question about tsdf HOT 3
- Reprojection yielded 0 valid points :( HOT 5
- Coordinate frame of point cloud expected HOT 1
- Generated surface is offset by a voxel width HOT 4
- Example data for tests HOT 9
- integrateCloud segmentation faullt HOT 10
- cpu_tsdf cannot compile with pcl 1.8.1 HOT 8
- How to get the pose.txt.... HOT 2
- wrong include path HOT 2
- pcd file format HOT 8
- use of normal vectors HOT 3
- Linker Error with boost::program_options (Boost 1.57.0 & PCL 1.9.0) HOT 1
- build with pcl 1.8.1 failure HOT 6
- is the hpp file:tsdf_volume_octree.hpp any useful in this project? HOT 1
- complied error with pcl1.10
- This project depend on which version of pcl HOT 1
- Can't generate tsdf2mesh executable when I switch pcl version to pcl 1.10 HOT 1
- Core dumped when I run integrate HOT 1
- rtabmap+cputsdf+pcl1.8.1 HOT 2
- How to run my local data? HOT 12
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from cpu_tsdf.