3 ms·

> new technology will enable … what exactly? Covered in the article > Our model is able to measure canopy coverage using newer lidar data than existing canop

> new technology will enable … what exactly? Covered in the article > Our model is able to measure canopy coverage using newer lidar data than existing canopy coverage maps and measure the size and volume of individual canopies, making it easier to see where healthy trees are located and where existing canopies are lacking

asdff 3y ago
I wish these sorts of studies quantified how much more useful they are than the last model. There's a lot of papers out there where they reproduce some finding from an older method or dataset, but use the new hot method or dataset. Very few of them go ahead and do power comparisons between old and new, to see if it was even worth all the trouble and by how much.
- cantSpellSober 3y ago
  I found this: https://labs.aap.cornell.edu/daslab/projects/treefolio https://labs.aap.cornell.edu/daslab/projects/treefolio The videos were ok > Tree Folio uses bespoke tools to identify tree canopies from raw LiDAR data. We've built a filtering program to isolate tree canopy points from the rest of the LiDAR data, a clustering algorithm to extract 3D models of each individual tree from the canopy, and a custom shading simulation platform that models shading interactions between the specific tree and the surrounding buildings
  - asdff 3y ago
    I went through their website, but it didn't really answer my question of if this was really better than some old fashioned way, e.g. taking satellite imagery and measuring density of canopy and using some simple formula to estimate shade effect based on tree or orientation of the street perhaps. 3D modelling, constantly updating LIDAR maps, custom shading simulation, it all seems like a lot of resource heavy technology. I'm not sure whether or not its being applied in a way that actually offers a statistically significant benefit over other, maybe less expensive methods.

asdff 3y ago

I wish these sorts of studies quantified how much more useful they are than the last model. There's a lot of papers out there where they reproduce some finding from an older method or dataset, but use the new hot method or dataset. Very few of them go ahead and do power comparisons between old and new, to see if it was even worth all the trouble and by how much.

cantSpellSober 3y ago
I found this: https://labs.aap.cornell.edu/daslab/projects/treefolio https://labs.aap.cornell.edu/daslab/projects/treefolio The videos were ok > Tree Folio uses bespoke tools to identify tree canopies from raw LiDAR data. We've built a filtering program to isolate tree canopy points from the rest of the LiDAR data, a clustering algorithm to extract 3D models of each individual tree from the canopy, and a custom shading simulation platform that models shading interactions between the specific tree and the surrounding buildings
- asdff 3y ago
  I went through their website, but it didn't really answer my question of if this was really better than some old fashioned way, e.g. taking satellite imagery and measuring density of canopy and using some simple formula to estimate shade effect based on tree or orientation of the street perhaps. 3D modelling, constantly updating LIDAR maps, custom shading simulation, it all seems like a lot of resource heavy technology. I'm not sure whether or not its being applied in a way that actually offers a statistically significant benefit over other, maybe less expensive methods.