Urban Tree Detection Data

[Ly0n]

This looks like a dataset missing that meets the requirements of OpenForest:
https://github.com/jonathanventura/urban-tree-detection-data

[ArthurOuaknine]

Thanks for flagging this dataset!
To maintain consistency in the repository, I would need further information about the dataset.
If anyone would like to contribute to the repo, I will be happy to integrate it myself with the related information.
One may just fill the proposed template and push it on a pull request, I'll do the rest.
More details about the guidelines are provided in the README.md.

• dataset_name: name of the dataset.
• article_url: indicate the url to the article associated to the dataset.
• category: depending on the modality available in the dataset, please indicate one or several letters as following: I for inventories, G for ground-based recordings, A for aerial recordings, S for satellite recordings, M for maps. For more information, please refer to our article.
• year_recordings: each year of data recording or indicate a time series. Note that Unknown is a valid entry if the recording date is not available.
• dataset_size: indicate the number of trees, number of samples, number of maps with the appropriate order of magnitude, i.e k for thousands, M for millions, B for billions. You can also indicate the area covers by the provided data with an appropriate order of magnitude, either ha or km2.
• data: indicate each modality available in the dataset e.g. Aerial RGB, Multispectral, SAR, LiDAR PC and so on.
• spatial_resolution_or_precision: according to the same order than in the data attribute, indicate the associated spatial resolution or precision of each modality provided in the dataset. The measure unit depends on the modality, e.g. centimeters, meters, kilometers or number of points per meter squared for point clouds.
• time_series: indicate Yes or No if the dataset contains a time series, note that it must be consistant with the year_recordings attribute.
• potential_tasks: indicate the potential tasks that the dataset could be used for. ( eg. alignment, change detection, classification, instance segmentation, key-point detection, multi classification, object detection, object localization, regression, semantic segmentation).
• nb_classes: indicate the number of classes in the datasets if applicable, N/A otherwise.
• location: indicate the countries in which the dataset is localized.
• license: indicate the license application to the dataset. Note that the file will be tested before merging to ensure that the format and typos are respected.

I can also add NaNs if info are missing.
Thanks for your contribution!