dymaxionlabs/satproc
:artificial_satellite: Python library and CLI tools for processing geospatial imagery for ML
PythonApache-2.0
Issues
- 0
a specific polygon for aoi
#49 opened by Ro-devesa - 0
- 0
- 0
- 0
If `--step-size` is not specified, it should use the same `--size` (no overlapping)
#17 opened by munshkr - 0
Implement missing `--mask-type single` option on `extract_chips` and `make_masks`
#18 opened by munshkr - 0
Spatial filtering (median, gaussian)
#41 opened by munshkr - 0
- 1
- 1
Write integration and unit tests
#34 opened by munshkr - 0
Improve documentation
#25 opened by munshkr - 0
- 0
Integrate CI with GitHub Actions
#30 opened by munshkr - 2
Delete images with empty masks
#29 opened by Ro-devesa - 0
Compute statistics of images
#27 opened by munshkr - 0
- 0
Pin versions: upgrade to latest pyproj (>=3.2)
#24 opened by munshkr - 0
On `polygonize` add option for extracting contours with marching cubes (`gdal_contour`) from probability/distance mask
#23 opened by munshkr - 0
In `polygonize`, try using `rasterio.features.shapes` to polygonize instead of `gdal_polygonize.py`
#20 opened by munshkr - 0
Create new module and CLI script `generalize` for simplyfing and smoothing polygon vector files
#22 opened by munshkr - 1
- 0
- 0
When using `whole=True`, `extract_chips` does not extract chips at borders if image size is not multiple of chip size
#4 opened by munshkr - 0
- 0
If chip size is bigger than some image size, it ends up writing a rescaled image
#16 opened by munshkr - 1
suggestion
#2 opened by rbavery - 1
- 0
Check and validate CRS when using `--aoi` and/or `--labels` with respect to CRS of images
#8 opened by munshkr - 0
`extract_chips` does not give good feedback to user when not extracting chips because of the "low contrast" check
#11 opened by munshkr - 0
By default always write chips, even if they do not pass the high contrast check
#10 opened by munshkr - 0
Improve documentation
#13 opened by munshkr - 0
- 0
- 0
Allow rescaling in different modes per band
#1 opened by munshkr - 1