Martin Valgur
Martin Valgur
Thank you for open sourcing the model! It's much appreciated. This PR adds a script to generate the surface normals ground truth required for training, which is currently not provided...
I noticed the camera intrinsics f, cx, and cy used in the [`surface-normals`](https://github.com/JiaxiongQ/surface-normal/blob/master/tool.cpp#L360) tool and in [`trainLoader.py`](https://github.com/JiaxiongQ/DeepLiDAR/blob/a19f04efad88e11cfdabb18a14765046f871972a/dataloader/trainLoader.py#L18) differ from the ones provided with KITTI for many but not all values....
Hi! This PR tweaks the file fixture loading in tests to use [`importlib_resources`](https://pypi.org/project/importlib-resources/) to ensure that tests can be run from any working directory (or even as part of a...
This makes initializing the classes a bit more convenient. The `cv2` dependency is quite heavy-weight and entirely unused right now. By the way, would it be possible to create a...
This PR touches nearly all parts of the code more or less. The main changes are: - Cleaner separation of raw and odom code. Moved all raw/odom specific code out...
Unlike other rinex types, SP3 files are currently not opened with `georinex.rio.opener()`, which means compressed and file-handle inputs are not handled as expected. This PR fixes that inconsistency.
Hi! This PR replaces `unlzw3` and `unlzw` with `ncompress` for LZW decompression. Support for LZW compression and decompression was a bit lacking in Python, so I created a new [ncompress](https://github.com/valgur/ncompress)...
Hi! This PR makes use of the [hatanaka](https://github.com/valgur/hatanaka) and [ncompress](https://github.com/valgur/ncompress) libraries for a more streamlined and robust decompression of downloaded RINEX products. Both are available from PyPI and support all...
A plugin for [Fleep](https://fleep.io). I'm not related to the authors of that application. Third-party auth is currently not working, but that seems to be a general problem in Franz at...
Hi, Thank you for releasing D-LIOM. I found the approach described in the paper to be very elegant and with exceptionally good reported performance, so I wanted to see if...