Last year I applied the algorithm used by the TERRA REF team to mask these images and calculate the percent of the image that is plant vs not plant (called canopy cover, a %).
\nNote: there should be one these may not exactly match
\nThe image masks identify which part of the image is plant, and which part is not plant, and are the same size as the input image.
\nIf values of canopy cover would be useful for this challenge, these values can be found in the compressed CSV file sorghum_biomass_canopycover.zip
\nThe CSV has two columns, the image name and the percent canopy cover.
\nIf masks would be useful, you should be able to run the code at https://github.com/cct-datascience/sorghum_biomass_prediction/. If you have trouble doing so, please let me know.
", "messages": [ { "id": "1753524", "postDate": "04/12/2022 23:32:30", "content": "Last year I applied the algorithm used by the TERRA REF team to mask these images and calculate the percent of the image that is plant vs not plant (called canopy cover, a %).
\nNote: there should be one these may not exactly match
\nThe image masks identify which part of the image is plant, and which part is not plant, and are the same size as the input image.
\nIf values of canopy cover would be useful for this challenge, these values can be found in the compressed CSV file sorghum_biomass_canopycover.zip
\nThe CSV has two columns, the image name and the percent canopy cover.
\nIf masks would be useful, you should be able to run the code at https://github.com/cct-datascience/sorghum_biomass_prediction/. If you have trouble doing so, please let me know.
", "rawMarkdown": "Last year I applied the algorithm used by the TERRA REF team to mask these images and calculate the percent of the image that is plant vs not plant (called canopy cover, a %). \n\nNote: there should be one these may not exactly match \n\nThe image masks identify which part of the image is plant, and which part is not plant, and are the same size as the input image.\n\nIf values of canopy cover would be useful for this challenge, these values can be found in the compressed CSV file [sorghum_biomass_canopycover.zip](https://github.com/cct-datascience/sorghum_biomass_prediction/releases/download/v1/sorghum_biomass_canopycover.zip)\n\nThe CSV has two columns, the image name and the percent canopy cover.\n\nIf masks would be useful, you should be able to run the code at https://github.com/cct-datascience/sorghum_biomass_prediction/. If you have trouble doing so, please let me know.", "votes": null }, { "id": "1759400", "postDate": "04/18/2022 15:37:48", "content": "Any way to generate mask on a kaggle notebook ? getting path errors
", "rawMarkdown": "Any way to generate mask on a kaggle notebook ? getting path errors", "votes": null }, { "id": "1759545", "postDate": "04/18/2022 18:12:26", "content": "I am not sure, I've never used a Kaggle notebook. What are the path errors that you are getting?
", "rawMarkdown": "I am not sure, I've never used a Kaggle notebook. What are the path errors that you are getting?", "votes": null }, { "id": "1760327", "postDate": "04/19/2022 08:07:27", "content": "Oh I found the error ; you cant run docker on kaggle notebook . It would be really helpful if you share the mask in a kaggle dataset
", "rawMarkdown": "Oh I found the error ; you cant run docker on kaggle notebook . It would be really helpful if you share the mask in a kaggle dataset", "votes": null }, { "id": "1769707", "postDate": "04/27/2022 14:07:50", "content": "Hello.
\nThanks for sharing the cool method, @dlebauer!
\nI tried to create a mask for the sorghum-fgvc9 train: 22193, test: 23639 image using your GitHub code and docker image.
\nIf I can get permission from @dlebauer, I would like to register these in the Public
\nKaggle dataset. I'd be glad to get your feedback!
Thank you for doing this. You are welcome to share. Since the data is public domain and the code is open source you don’t actually need my permission, but citation in any publications would be appreciated:
\nto cite the Sorghum 100 dataset:
\n\n\nRen, C., Dulay, J., Rolwes, G., Pauli, D., Shakoor, N., & Stylianou, A. (2021). Multi-resolution outlier pooling for sorghum classification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 2931-2939).
\n
to cite the masking algorithm:
\n\n\nBurnette et al (2019) terraref/extractors-stereo-rgb: Season 6 Data Publication (2019) (Version S6_Pub_2019). Zenodo. http://doi.org/10.5281/zenodo.3406304
\n
to cite the workflow:
\n\n", "rawMarkdown": "Thank you for doing this. You are welcome to share. Since the data is public domain and the code is open source you don’t actually need my permission, but citation in any publications would be appreciated:\n\n---\n\nto cite the Sorghum 100 dataset:\n> Ren, C., Dulay, J., Rolwes, G., Pauli, D., Shakoor, N., & Stylianou, A. (2021). Multi-resolution outlier pooling for sorghum classification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 2931-2939).\n\nto cite the masking algorithm: \n> Burnette et al (2019) terraref/extractors-stereo-rgb: Season 6 Data Publication (2019) (Version S6_Pub_2019). Zenodo. http://doi.org/10.5281/zenodo.3406304\n\nto cite the workflow: \n> Schnaufer, C., & LeBauer, D. (2022) Generate Image Masks and Canopy Cover from Sorghum 100 Images [Computer software]. https://github.com/cct-datascience/sorghum_100_masks https://doi.org/10.5281/zenodo.6456476", "votes": null }, { "id": "1770617", "postDate": "04/28/2022 12:30:35", "content": "Schnaufer, C., & LeBauer, D. (2022) Generate Image Masks and Canopy Cover from Sorghum 100 Images [Computer software]. https://github.com/cct-datascience/sorghum_100_masks https://doi.org/10.5281/zenodo.6456476
\n
Thanks for the reply, and for the advice!
\nThe mask image is now available at the following URL.
\nI have clearly stated the algorithm and the source of the data.
\nhttps://www.kaggle.com/datasets/bobfromjapan/sorghumidfgvc9masking
If you see any problems, please comment!
", "rawMarkdown": "Thanks for the reply, and for the advice!\n\nThe mask image is now available at the following URL.\nI have clearly stated the algorithm and the source of the data.\nhttps://www.kaggle.com/datasets/bobfromjapan/sorghumidfgvc9masking\n\nIf you see any problems, please comment!", "votes": null } ], "comments": [ { "id": 1759400, "author_name": "mithilsalunkhe", "author_url": "", "post_date": "04/18/2022 15:37:48", "content": "Any way to generate mask on a kaggle notebook ? getting path errors
", "votes": null, "replies": [ { "id": 1759545, "author_name": "dlebauer", "author_url": "", "post_date": "04/18/2022 18:12:26", "content": "I am not sure, I've never used a Kaggle notebook. What are the path errors that you are getting?
", "votes": null, "replies": [] }, { "id": 1760327, "author_name": "mithilsalunkhe", "author_url": "", "post_date": "04/19/2022 08:07:27", "content": "Oh I found the error ; you cant run docker on kaggle notebook . It would be really helpful if you share the mask in a kaggle dataset
", "votes": null, "replies": [] }, { "id": 1769707, "author_name": "bobfromjapan", "author_url": "", "post_date": "04/27/2022 14:07:50", "content": "Hello.
\nThanks for sharing the cool method, @dlebauer!
\nI tried to create a mask for the sorghum-fgvc9 train: 22193, test: 23639 image using your GitHub code and docker image.
\nIf I can get permission from @dlebauer, I would like to register these in the Public
\nKaggle dataset. I'd be glad to get your feedback!
Thank you for doing this. You are welcome to share. Since the data is public domain and the code is open source you don’t actually need my permission, but citation in any publications would be appreciated:
\nto cite the Sorghum 100 dataset:
\n\n\nRen, C., Dulay, J., Rolwes, G., Pauli, D., Shakoor, N., & Stylianou, A. (2021). Multi-resolution outlier pooling for sorghum classification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 2931-2939).
\n
to cite the masking algorithm:
\n\n\nBurnette et al (2019) terraref/extractors-stereo-rgb: Season 6 Data Publication (2019) (Version S6_Pub_2019). Zenodo. http://doi.org/10.5281/zenodo.3406304
\n
to cite the workflow:
\n\n", "votes": null, "replies": [] }, { "id": 1770617, "author_name": "bobfromjapan", "author_url": "", "post_date": "04/28/2022 12:30:35", "content": "Schnaufer, C., & LeBauer, D. (2022) Generate Image Masks and Canopy Cover from Sorghum 100 Images [Computer software]. https://github.com/cct-datascience/sorghum_100_masks https://doi.org/10.5281/zenodo.6456476
\n
Thanks for the reply, and for the advice!
\nThe mask image is now available at the following URL.
\nI have clearly stated the algorithm and the source of the data.
\nhttps://www.kaggle.com/datasets/bobfromjapan/sorghumidfgvc9masking
If you see any problems, please comment!
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1753524": "Last year I applied the algorithm used by the TERRA REF team to mask these images and calculate the percent of the image that is plant vs not plant (called canopy cover, a %). \n\nNote: there should be one these may not exactly match \n\nThe image masks identify which part of the image is plant, and which part is not plant, and are the same size as the input image.\n\nIf values of canopy cover would be useful for this challenge, these values can be found in the compressed CSV file [sorghum_biomass_canopycover.zip](https://github.com/cct-datascience/sorghum_biomass_prediction/releases/download/v1/sorghum_biomass_canopycover.zip)\n\nThe CSV has two columns, the image name and the percent canopy cover.\n\nIf masks would be useful, you should be able to run the code at https://github.com/cct-datascience/sorghum_biomass_prediction/. If you have trouble doing so, please let me know.", "1759400": "Any way to generate mask on a kaggle notebook ? getting path errors", "1759545": "I am not sure, I've never used a Kaggle notebook. What are the path errors that you are getting?", "1760327": "Oh I found the error ; you cant run docker on kaggle notebook . It would be really helpful if you share the mask in a kaggle dataset", "1769707": "Hello.\n\nThanks for sharing the cool method, @dlebauer! \nI tried to create a mask for the sorghum-fgvc9 train: 22193, test: 23639 image using your GitHub code and docker image.\nIf I can get permission from @dlebauer, I would like to register these in the Public\nKaggle dataset. I'd be glad to get your feedback!", "1769774": "Thank you for doing this. You are welcome to share. Since the data is public domain and the code is open source you don’t actually need my permission, but citation in any publications would be appreciated:\n\n---\n\nto cite the Sorghum 100 dataset:\n> Ren, C., Dulay, J., Rolwes, G., Pauli, D., Shakoor, N., & Stylianou, A. (2021). Multi-resolution outlier pooling for sorghum classification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 2931-2939).\n\nto cite the masking algorithm: \n> Burnette et al (2019) terraref/extractors-stereo-rgb: Season 6 Data Publication (2019) (Version S6_Pub_2019). Zenodo. http://doi.org/10.5281/zenodo.3406304\n\nto cite the workflow: \n> Schnaufer, C., & LeBauer, D. (2022) Generate Image Masks and Canopy Cover from Sorghum 100 Images [Computer software]. https://github.com/cct-datascience/sorghum_100_masks https://doi.org/10.5281/zenodo.6456476", "1770617": "Thanks for the reply, and for the advice!\n\nThe mask image is now available at the following URL.\nI have clearly stated the algorithm and the source of the data.\nhttps://www.kaggle.com/datasets/bobfromjapan/sorghumidfgvc9masking\n\nIf you see any problems, please comment!" }, "source": "meta" }