From f6c82035f5591a58aa0e862c2675308a10e875f9 Mon Sep 17 00:00:00 2001 From: jacquelynsmale Date: Thu, 6 Nov 2025 16:10:55 -0900 Subject: [PATCH 1/2] make some small tweaks --- .../generate-chips/generating_chip_data.md | 16 ++++++++-------- .../training/run_severe_weather_segmentation.md | 2 +- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/severe-weather/generate-chips/generating_chip_data.md b/severe-weather/generate-chips/generating_chip_data.md index c805fda..bce385f 100644 --- a/severe-weather/generate-chips/generating_chip_data.md +++ b/severe-weather/generate-chips/generating_chip_data.md @@ -6,7 +6,7 @@ ## 1. Create Conda Env -All the dependencies for these scripts are in one conda environment, which can be installed and activated with: +All the dependencies for these scripts are in the `severe-weather` conda environment, which can be installed and activated with: ``` conda env create -f environment.yml @@ -15,28 +15,28 @@ conda activate severe-weather ## 2. Make Label Rasters -With the conda env activated the rasters can be generated by running: +With the `severe-weather` conda environment activated, the rasters can be generated by running: ``` python generate_severe_weather_rasters.py ``` -This script downloads the [hwds_db](https://github.com/jrbell1/hwds_db) dataset and turns all the -damasge swaths during Sentinel 1 into 10m rasters. +This script downloads the [Hail and Wind Damage Swath (HWDS)](https://github.com/jrbell1/hwds_db) dataset and turns all the +swaths during Sentinel 1 into 10m rasters. ## 3. Generate Label Chips from Rasters -To generate label and data chips from the hwds rasters run: +To generate label and data chips from the hwds rasters, run: ``` python chip_severe_weather_raster.py (tif path) ``` -The tifs can be found in the `./data/hwds-rasters/` directory made by `generate_severe_weather_rasters.py`. -One example command would be: +The tifs can be found in the `data/hwds-rasters/` directory made by `generate_hwds_rasters.py`. +For example, you could generate files with the swath ID 1507 acquired on July 6, 2020 with: ``` python chip_severe_weather_raster.py data/severe-weather-rasters/swathID_1507_swathDate_2020-07-06.tif ``` -This will generate zarr files for each chip and in the `chips` directory. There will be a directory for the labels, one for the S1RTC chips and one for S2L2A chips. +This will generate 3 zarr files in the `data/zarrs` directory: one for the labeled data (`swathID_1507_swathDate_2020-07-06.zarr.zip`), one for the S1RTC chips (`swathID_1507_swathDate_2020-07-06_S1RTC.zarr.zip`), and one for S2L2A chips (`swathID_1507_swathDate_2020-07-06_S2L2A.zarr.zip`). diff --git a/severe-weather/training/run_severe_weather_segmentation.md b/severe-weather/training/run_severe_weather_segmentation.md index 962b2e4..d6aefcd 100644 --- a/severe-weather/training/run_severe_weather_segmentation.md +++ b/severe-weather/training/run_severe_weather_segmentation.md @@ -9,7 +9,7 @@ The data are in `zarr` format and prepared using [SatChip](https://github.com/fo * `S1RTC`: Sentinel-1 Radiometric Terrain Corrected (RTC) data created using [ASF's HyP3 on-demand platform](https://hyp3-docs.asf.alaska.edu/guides/rtc_product_guide/) * `HYP3S1RTC`: Sentinel-1 Radiometric Terrain Corrected (RTC) data created using [ASF's HyP3 on-demand platform](https://hyp3-docs.asf.alaska.edu/guides/rtc_product_guide/) -SatChip prepares data labels and satellite images into 264x264 sample files that follow the TerraMind extension of the MajorTom specification and are saved as `zarr.zip` files. [Here](generating_chip_data.md) is a tutorial on generating your own chip data. +SatChip prepares data labels and satellite images into 264x264 image arrays that follow the TerraMind extension of the MajorTom specification and are saved as `zarr.zip` files. You can walk through [this tutorial](generating_chip_data.md) to generate your own chip data. Below is an example of metadata associated with the Sentinel-1 RTC chip `swathID_1260_swathDate_2019-06-17_464U_867L_2_0_S1RTC.zarr.zip`: ``` From 64fdccf243d99a80f0ea25a8e3b238e3f6d92792 Mon Sep 17 00:00:00 2001 From: Jacquelyn Smale <34557291+jacquelynsmale@users.noreply.github.com> Date: Fri, 7 Nov 2025 09:25:07 -0900 Subject: [PATCH 2/2] Update severe-weather/generate-chips/generating_chip_data.md Co-authored-by: William Horn --- severe-weather/generate-chips/generating_chip_data.md | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/severe-weather/generate-chips/generating_chip_data.md b/severe-weather/generate-chips/generating_chip_data.md index bce385f..b1e7523 100644 --- a/severe-weather/generate-chips/generating_chip_data.md +++ b/severe-weather/generate-chips/generating_chip_data.md @@ -21,8 +21,7 @@ With the `severe-weather` conda environment activated, the rasters can be genera python generate_severe_weather_rasters.py ``` -This script downloads the [Hail and Wind Damage Swath (HWDS)](https://github.com/jrbell1/hwds_db) dataset and turns all the -swaths during Sentinel 1 into 10m rasters. +This script downloads the [Hail and Wind Damage Swath (HWDS)](https://github.com/jrbell1/hwds_db) dataset and takes all the damage swaths that occurred during Sentinel 1 mission (2015-Present) and turns them into 10m label rasters. ## 3. Generate Label Chips from Rasters