Vegetation mapping dataset of Gran Canaria, Canary Islands, Spain.

Description

This dataset consists primarily of satellite images of the island of Gran Canaria and masks that segment the vegetation communities present in them.The original satellite images of the island are extracted from open data provided by the Regional Government through the company GRAFCAN. The company also has a visor that exemplify the quality of the images (Link).

The dataset contains 185 images for 2016, 2018, 2021, summing up a total of 555 images in 185 positions. In this way we are able to see the change through the years.

Additionally to the mask that segmentate the communities, each image, the dataset also have the altitude data and, only for 2018 and 2021, the infrared data as is capture by the GRAFCAN cameras.

Files description

Sat_images: Contains the RGB satellite images of the locations annotated.     Is divided in 3 years, each with the images of that location on that year.

Masks: Contains the RGB masks with the annotations of each image each year. Is divided in    the base version, which only has information about the vegetation aggrouping all non vegetation structures as the same class    and the detailed version, which has different annotations for bodies of water, roads...

Additional_data: Contains Additional information of the images, the altitude data of the location in meters    and the infrared images, these images are only available for 2018 and 2021, since they were not recorded in 2016.

CSV_file: Contains the csv files with all kinds of information about the images. Next, we proceed to detail each file.

Name_Translations.csv and Name_Translations_Detailed.csv : In the other files the classes names are always in spanish, as some vegetation communities do not have a direct or easy translation to english.

Colors.csv and Colors_detailed.csv: csv files that contain the RGB values and HEX codes of the colors assigned to the different classes.

Annotations.csv: csv file with information about how many communities are in each image and which ones are. If the versions of one image, for example, GC_011 has variations over the years in the communities, an extra entry marked as GC_011_2018 or GC_011_2021 appears at the end of csv with the new data of the image in that year. If it is marked in 2018, it means that this change also happens in 2021.

Detailed_annotations : Same as before, they are splitted because there are much more differences in appearance than in the base annotations.

Coordinates.csv: This csv file contains the GEOtransformation that crop the image from the original .jp2 map. Inside this, there is the information of the top-left corner of the image, meaning that it is possible to re-capture all the images.

test_partition.txt: It contains a list of images that were used as the test partition in the experiments conducted

Citing the dataset

This dataset is released under a Creative Commons Attribution license, so please cite it if used in your work in any form. Additionally, this dataset is linked to two academic papers that are currently unpublished. Published academic papers should revisit this page in the future to be able to cite the two papers right. Personal works, such as personal projects or blog posts, should provide a URL to this Zenodo page, though a reference to the main papers would also be appreciated.

Contact information

If you would like further information about the dataset or if you experience any issues with the files, please contact at [e-mail]