In the CSIRO Image2Biomass Prediction dataset, the provided values for each image represent the biomass within a 70×30 cm quadrat.\nMy questions:\nAre these values biomass densities (kg/m²) or absolute biomass (kg) within the quadrat?\nIf they are absolute biomass, what is the recommended approach to estimate total biomass for a full hectare accurately without over-simplifying?\nThanks in advance for any guidance or insights!
", "messages": [ { "id": "3378247", "postDate": "12/17/2025 19:51:31", "content": "In the CSIRO Image2Biomass Prediction dataset, the provided values for each image represent the biomass within a 70×30 cm quadrat.\nMy questions:\nAre these values biomass densities (kg/m²) or absolute biomass (kg) within the quadrat?\nIf they are absolute biomass, what is the recommended approach to estimate total biomass for a full hectare accurately without over-simplifying?\nThanks in advance for any guidance or insights!
", "rawMarkdown": "In the CSIRO Image2Biomass Prediction dataset, the provided values for each image represent the biomass within a 70×30 cm quadrat.\n\nMy questions:\n\nAre these values biomass densities (kg/m²) or absolute biomass (kg) within the quadrat?\n\nIf they are absolute biomass, what is the recommended approach to estimate total biomass for a full hectare accurately without over-simplifying?\n\nThanks in advance for any guidance or insights!", "votes": null }, { "id": "3378352", "postDate": "12/17/2025 23:51:30", "content": "The provided values for each image represent the absolute biomass (in grams) within the 70×30 cm quadrat. You can also consider it as the average density over a small area, since the area corresponding to each image is consistent.
\nAlthough our ultimate scope involves handling very large areas, we cannot directly collect ground truth biomass data for such large expanses. Therefore, we must start with smaller areas. Predictions based on ground-level images can provide calibration for large-scale biomass predictions using satellite/aerial NDVI/visible light imagery. Additionally, we can use large-scale pasture image segmentation prediction to calculate sums, or train on larger-scale images by stitching and fusing them to form extensive images and total biomass estimates. If you are interested in this field, I would be happy to further communicate after the competition ends.
", "rawMarkdown": "The provided values for each image represent the absolute biomass (in grams) within the 70×30 cm quadrat. You can also consider it as the average density over a small area, since the area corresponding to each image is consistent.\n\nAlthough our ultimate scope involves handling very large areas, we cannot directly collect ground truth biomass data for such large expanses. Therefore, we must start with smaller areas. Predictions based on ground-level images can provide calibration for large-scale biomass predictions using satellite/aerial NDVI/visible light imagery. Additionally, we can use large-scale pasture image segmentation prediction to calculate sums, or train on larger-scale images by stitching and fusing them to form extensive images and total biomass estimates. If you are interested in this field, I would be happy to further communicate after the competition ends.", "votes": null }, { "id": "3378513", "postDate": "12/18/2025 12:16:23", "content": "Hi Qiyu,
\nI’m very interested in biomass prediction and the idea of scaling from quadrat-level data to hectare-scale estimates. I would love to explore approaches for extending predictions to a full hectare and integrating with aerial or satellite imagery.
\nI’d be happy to discuss this further after the competition if you’re open to it
", "rawMarkdown": "Hi Qiyu,\n\nI’m very interested in biomass prediction and the idea of scaling from quadrat-level data to hectare-scale estimates. I would love to explore approaches for extending predictions to a full hectare and integrating with aerial or satellite imagery.\n\nI’d be happy to discuss this further after the competition if you’re open to it", "votes": null } ], "comments": [ { "id": 3378352, "author_name": "qiyuliaocsiro", "author_url": "", "post_date": "12/17/2025 23:51:30", "content": "The provided values for each image represent the absolute biomass (in grams) within the 70×30 cm quadrat. You can also consider it as the average density over a small area, since the area corresponding to each image is consistent.
\nAlthough our ultimate scope involves handling very large areas, we cannot directly collect ground truth biomass data for such large expanses. Therefore, we must start with smaller areas. Predictions based on ground-level images can provide calibration for large-scale biomass predictions using satellite/aerial NDVI/visible light imagery. Additionally, we can use large-scale pasture image segmentation prediction to calculate sums, or train on larger-scale images by stitching and fusing them to form extensive images and total biomass estimates. If you are interested in this field, I would be happy to further communicate after the competition ends.
", "votes": null, "replies": [ { "id": 3378513, "author_name": "samad23", "author_url": "", "post_date": "12/18/2025 12:16:23", "content": "Hi Qiyu,
\nI’m very interested in biomass prediction and the idea of scaling from quadrat-level data to hectare-scale estimates. I would love to explore approaches for extending predictions to a full hectare and integrating with aerial or satellite imagery.
\nI’d be happy to discuss this further after the competition if you’re open to it
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "3378247": "In the CSIRO Image2Biomass Prediction dataset, the provided values for each image represent the biomass within a 70×30 cm quadrat.\n\nMy questions:\n\nAre these values biomass densities (kg/m²) or absolute biomass (kg) within the quadrat?\n\nIf they are absolute biomass, what is the recommended approach to estimate total biomass for a full hectare accurately without over-simplifying?\n\nThanks in advance for any guidance or insights!", "3378352": "The provided values for each image represent the absolute biomass (in grams) within the 70×30 cm quadrat. You can also consider it as the average density over a small area, since the area corresponding to each image is consistent.\n\nAlthough our ultimate scope involves handling very large areas, we cannot directly collect ground truth biomass data for such large expanses. Therefore, we must start with smaller areas. Predictions based on ground-level images can provide calibration for large-scale biomass predictions using satellite/aerial NDVI/visible light imagery. Additionally, we can use large-scale pasture image segmentation prediction to calculate sums, or train on larger-scale images by stitching and fusing them to form extensive images and total biomass estimates. If you are interested in this field, I would be happy to further communicate after the competition ends.", "3378513": "Hi Qiyu,\n\nI’m very interested in biomass prediction and the idea of scaling from quadrat-level data to hectare-scale estimates. I would love to explore approaches for extending predictions to a full hectare and integrating with aerial or satellite imagery.\n\nI’d be happy to discuss this further after the competition if you’re open to it" }, "source": "meta" }