Multi-temporal 3D Models of Antarctic Grounded Ice Cliff for Calving Monitoring

(For the rendered version, please see README.pdf)
Dataset
This is the dataset of the paper "Fine-scale Antarctic Grounded Ice Cliff 3D Calving Monitoring based on Multi-temporal UAV Photogrammetry Without Ground Control". The dataset contains geo-registrated multitemporal meshes recording the calving process of an ice cliff in Antarctica and ice cliff calving detection results.

Multitemporal meshes
The meshes were generated from UAV photogrammetry and co-alignment technique. The dataset is a single CloudCompare `.bin` file, `Ice_cliff_multitemporal_mesh.bin`. It contains meshes generated by Metashape from each epoch. The file structure is shown in the below image. Each subfolder contains a mesh of the ice cliff at `yyyymmdd` day.

The ice cliff is located between the Qinling station area (74°56' S, 163° 42' E) and the Nansen Ice shelf, Inexpressive Island in Victoria Land, East Antarctica. The length of its coastline is approximately 0.89 km.

The UAV for collecting aerial images was a DJI Mavic 2 Pro drone, which carries a Hasselblad L1D-20c gimbal camera. Its 28 mm-equivalent lens has a 77° field of view (FoV). From 30-Jan
to 25-Feb 2022, 10 flights were performed with the same flight parameters in 26 days. The date of flights were 30-Jan, 31-Jan, 1-Feb, 5-Feb, 7-Feb, 10-Feb, 13-Feb, 16-Feb, 17-Feb and 25-Feb.

|Parameter name| Value|
|----------------------| -----------------------------------
Flying height | 100 m
Image front/sidelap | 80% / 60%
\# of images planned | 330
GSD | 2.1 cm
Image coverage | 159 $\times$ 106 m

Calving detection results
In the `\result` folder stores a series of point cloud files `mmdd-mmdd_diff.ply`. They are calving object detection results of our paper.