Hail Event on 2021-06-20 in Entlebuch (LU), Switzerland: Drone Photogrammetry Imagery, Hail Sensor Recordings, Mask R-CNN Model and Analysis Data of Hailstones
Contributors
Data collectors:
- 1. Federal Office of Meteorology and Climatology, MeteoSwiss
- 2. ETH Zurich
- 3. Oeschger Centre for Climate Change Research and Institute of Geography, University of Bern, Bern, Switzerland
Description
This data collection belongs to the AMT publication https://amt.copernicus.org/preprints/amt-2023-89/ with the title "Drone-based photogrammetry combined with deep-learning to estimate hail size distributions and melting of hail on the ground".
Abstract:
Hail is a major threat associated with severe thunderstorms and estimating the hail size is important for issuing warnings to the public. For the validation of existing, operational, radar-derived hail estimates, ground-based observations are necessary. Automatic hail sensors, as for example within the Swiss hail network, record the kinetic energy of hailstones to estimate the hail sizes. Due to the small size of the observational area of these sensors (0.2 m2), the full hail size distribution (HSD) cannot be retrieved. To address this issue, we apply a state-of-the-art custom trained deep-learning object detection model to drone-based aerial photogrammetric data to identify hailstones and estimate the HSD. Photogrammetric data of hail on the ground was collected for one supercell thunderstorm crossing central Switzerland from southwest to northeast in the afternoon of 20 June 2021. The hail swath of this intense right-moving supercell was intercepted a few minutes after the passage at a soccer field near Entlebuch (Canton Lucerne, Switzerland) and aerial images were taken by a commercial DJI drone, equipped with a 45 megapixel full frame camera system. The resulting images have a ground sampling distance (GSD) of 1.5 mm per pixel, defined by the focal length of 35 mm of the camera and a flight altitude of 12 m above ground. A 2D orthomosaic model of the survey area (750.4 m2) is created based on 116 captured images during the first drone mapping flight. Hail is then detected by using a region-based Convolutional Neural Network (Mask R-CNN). We first characterize the hail sizes based on the individual hail segmentation masks resulting from the model detections and investigate the performance by using manual hail annotations by experts to generate validation and test data sets. The final HSD, composed of 18207 hailstones, is compared with nearby automatic hail sensor observations, the operational weather radar based hail product MESHS (Maximum Expected Severe Hail Size) and crowdsourced hail reports. Based on the retrieved data set, a statistical assessment of sampling errors of hail sensors is carried out. Furthermore, five repetitions of the drone-based photogrammetry mission within 18.65 min facilitate investigations into the hail melting process on the ground.
Table of contents (English)
- Annotation_datasets
- Test_dataset
- Expert_A
- instances_default.json (Hailstone annotation data in the COCO JSON format)
- Expert_B
- instances_default.json (Hailstone annotation data in the COCO JSON format)
- Expert_C
- instances_default.json (Hailstone annotation data in the COCO JSON format)
- Images (All tile images 500 x 500px used in the test dataset)
- Expert_A
- Training_dataset
- Expert_A
- instances_default.json (Hailstone annotation data in the COCO JSON format)
- Images (All tile images 500 x 500px used in the training dataset)
- Expert_A
- Validation_dataset
- Expert_A
- instances_default.json (Hailstone annotation data in the COCO JSON format)
- Images (All tile images 500px x 500px used in the validation dataset)
- Expert_A
- Test_dataset
- Orthophotos
- Flight_01_FA_20210620.png (Full orthophoto of the hail survey area on 2021-06-20)
- Georeferencing (Information about georeferencing from the OpenDroneMap software)
- coords.txt
- georeferenced_model.boundary.json
- georeferenced_model.bounds.geojson
- georeferenced_model.bounds.gpkg
- georeferenced_model.info.json
- georeferenced_model.laz
- georeferenced_model.summary.json
- georeferencing_model_geo.txt
- proj.txt
- Melting
- Flight_01_MidC.png (Orthophoto of the soccer middle circle from 1. drone flight)
- Flight_02_MidC.png (Orthophoto of the soccer middle circle from 2. drone flight)
- Flight_03_MidC.png (Orthophoto of the soccer middle circle from 3. drone flight)
- Flight_04_MidC.png (Orthophoto of the soccer middle circle from 4. drone flight)
- Flight_05_MidC.png (Orthophoto of the soccer middle circle from 5. drone flight)
- Trained_model
- Evaluation (Data from model evaluations detectron2.evaluation.coco_evaluation)
- Test
- Expert_A
- coco_instances_results.json
- instances_predictions.pth
- Expert_B
- coco_instances_results.json
- instances_predictions.pth
- Expert_C
- coco_instances_results.json
- instances_predictions.pth
- Expert_A
- Validation
- coco_instances_results.json
- instances_predictions.pth
- Test
- State_dictionary
- model_final.pth (Final model state as serialized PyTorch state dictionary)
- Evaluation (Data from model evaluations detectron2.evaluation.coco_evaluation)
- Hail_size_data
- Flight_01_FA_20210620.nc (NetCDF file containing the hail size data from Flight_01_FA_20210620.png)
- Melting
- Flight_01_MidC.nc (NetCDF file containing the hail size data from Flight_01_MidC.png)
- Flight_02_MidC.nc (NetCDF file containing the hail size data from Flight_02_MidC.png)
- Flight_03_MidC.nc (NetCDF file containing the hail size data from Flight_03_MidC.png)
- Flight_04_MidC.nc (NetCDF file containing the hail size data from Flight_04_MidC.png)
- Flight_05_MidC.nc (NetCDF file containing the hail size data from Flight_05_MidC.png)
- Hail_sensor_recordings.csv (Automatic hail sensor records in the surrounding of the drone survey)
Technical info (English)
Example NetCDF header information for the Hail_size_data:
Flight_01_FA_20210620 {
dimensions:
n = 18212 ;
variables:
float gsd ;
gsd:description = "Ground Sampling Distance" ;
gsd:units = "mm/pixel" ;
float hail_minor_axis(n) ;
hail_minor_axis:description = "Hailstone minor axis length in the image plane" ;
hail_minor_axis:units = "mm" ;
float hail_major_axis(n) ;
hail_major_axis:description = "Hailstone major axis lenght in the image plane" ;
hail_major_axis:units = "mm" ;
int64 hail_center_lightness(n) ;
hail_center_lightness:description = "Lightness value of the center hailstone pixel in digital number (DN) ranging from 0-255 (8-bit image depth)" ;
hail_center_lightness:units = "DN" ;
int64 hail_center_x(n) ;
hail_center_x:description = "X-coordinate of the hailstone center pixel within the orthophoto" ;
hail_center_x:units = "pixels" ;
int64 hail_center_y(n) ;
hail_center_y:description = "Y-coordinate of the hailstone center pixel within the orthophoto" ;
hail_center_y:units = "pixels" ;
// global attributes:
:description = "NetCDF file containing hailstone data from a storm intercepted near Entlebuch (Canton Lucerne, Switzerland) on 2021-06-20. The data is retrieved from orthophoto Flight_01_FA_20210620.png" ;
:author = "Martin Lainer (martin.lainer@meteoswiss.ch)" ;
}
Notes (English)
We thank the Swiss insurance company La Mobilière for funding the installation and operation of the automatic hail sensor network and making the hail sensor data available.
Files
Annotation_datasets.zip
Files
(2.8 GB)
| Name | Size | Download all |
|---|---|---|
|
md5:d5c81454139dbfd70ccfeb5768080175
|
121.7 MB | Preview Download |
|
md5:e9eba8ca88452af8be142276bce0b211
|
4.5 kB | Preview Download |
|
md5:4748759978d21968706bfd93567b90c7
|
248.1 kB | Preview Download |
|
md5:d40ef9b286b069e6708d1cbe1afc8acc
|
2.4 GB | Preview Download |
|
md5:086d46fab7a7dae3f5d1d25c8f532e4c
|
328.1 MB | Preview Download |
Additional details
Related works
- Is supplement to
- Journal article: 10.5194/amt-2023-89 (DOI)
Dates
- Created
-
2024-02-20
- Collected
-
2021-06-20