Comments (3)
Just an update of what the robust estimator looks like (with relative timestamps on the y axis). I think we should be good!
from pit30m.
I downloaded the metadata files for the three frames above:
# off-place
In [3]: a = np.load("035046.meta.npy", allow_pickle=True)
In [4]: a
Out[4]:
array({'sequence_counter': 66215, 'gain_db': 0.0, 'transmission_seconds': Decimal('1194377401.350424528'), 'shutter_seconds': 0.000810178, 'capture_seconds': Decimal('1194377401.333333969'), 'image_channel': '/hdcam/12_middle_front_roof_wide/image'},
dtype=object)
# before
In [5]: a = np.load("005062.meta.npy", allow_pickle=True)
In [6]: a
Out[6]:
array({'sequence_counter': 66215, 'gain_db': 0.0, 'transmission_seconds': Decimal('1194377401.350424528'), 'shutter_seconds': 0.000810178, 'capture_seconds': Decimal('1194377401.333333969'), 'image_channel': '/hdcam/12_middle_front_roof_wide/image'},
dtype=object)
# after
In [7]: a = np.load("005063.meta.npy", allow_pickle=True)
In [8]: a
Out[8]:
array({'sequence_counter': 66216, 'gain_db': 0.0, 'transmission_seconds': Decimal('1194377401.451511383'), 'shutter_seconds': 0.0008101790000000001, 'capture_seconds': Decimal('1194377401.433331966'), 'image_channel': '/hdcam/12_middle_front_roof_wide/image'},
dtype=object)
It seems that, for some reason, frames 1 and 2 ended up with the exact same metadata.
from pit30m.
Small update on this:
- I've made more videos and visually confirmed that using the frame name as an index results in temporally-consistent frames
- I've made a smalls script that runs RANSAC on a function where x=frame_index, and y=timestamp, and observe that a slope of 0.1 fits the data almost perfectly, save for a few outliers that do seem to have obviously wrong timestamps.
- This is good news! It means we can automatically detect frames with corrupt timestamps and interpolate their correct timestamps from a robust estimator. I'll make a detailed PR soon.
from pit30m.
Related Issues (20)
- add partitioning logic
- Reproduce NetVLAD baseline
- Reproduce ResNet50 baselines
- Visualizations
- Ensure sensor indices are sorted by sensor timestamp
- Support submap gaps when indexing in order to cover the two currently unindexable logs
- Automate benchmarking of LogReader methods
- Convert everything to UNIX or GPS timestamps HOT 4
- Data optimization epic
- Unbreak tutorial_00_introduction
- Data cleaning epic
- optimize LiDAR to non-redundant vehicle-relative f32
- Remove "no GT available" README warning and release 0.0.2
- Noisy warnings in Torch example
- HD Map Missing? HOT 1
- Clarify data license and access rights in docs
- Add `__version__` to released packages
- How to access IMU/GPS data? HOT 2
- LiDAR -> IMU transformation HOT 3
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 pit30m.