Experiments related to things to do with SAR data that might be useful. Seeing some LLMs go with answers as a starting point - generally for things not boilerplate, not going to do too well. Even more so with science.
DO NOT assume any of the code here will run - any of it does, will note and put in a ROBOTWORKING folder. I will put anything I write into the src folder.
- Deriving Surface Resistivity from Polarimetric SARData Using Dual-Input UNet : https://arxiv.org/pdf/2207.01811.pdf
- AWS requester pays https://registry.opendata.aws/sentinel-1/
- US only https://scottyhq.github.io/sentinel1-rtc-stac/#/
- AWS catalogue https://github.com/awslabs/open-data-registry/blob/main/datasets/sentinel-1-rtc-indigo.yaml
- DEA only has a few isolated datasets - e.g. some NT ports etc.
- https://geoscience.data.qld.gov.au/data/report/cr124399
- otherwise known as first survey I thought of online
- https://github.com/RichardScottOZ/openSAR
- https://github.com/Narayana-Rao/PolSAR-tools [QGIS plugin mentioned in paper]
- https://github.com/Narayana-Rao/polsartools
- https://github.com/aalling93/Sentinel_1_python
- Can ChatGPT 3.5 lift an architecture out of a paper
- likely yes
- get it right, very little chance
- 5 tries to correct an error to get an actual 'model'
- Not a correct model, but a model that doesn't have actual errors
- Produces a different code version
- Same result of course
- Gets stuck in a loop trying to fix a different error
- https://www.phind.com/search?cache=wlvfrv0bddld5xahaeyko2j1
- Perplexity free
- Would expect same things
- Produced a no-error 'model' first try
- https://www.perplexity.ai/search/Please-turn-this-SeRVbVKQTlWMcZQSRCcong
- Most robots can do things like turn text into tables and perplexity got this first go
- Asked for code there - get a generic requests request - as it doesn't actually know the api
- Vaguely understands this - but only in boilerplate requests fashion, as you would expect
| Short Form | Description |
|---|---|
| σ◦HH | Backscatter intensity (HH polarization) |
| σ◦HV | Backscatter intensity (HV polarization) |
| σ◦VV | Backscatter intensity (VV polarization) |
| λ1, λ2, λ3 | Eigen values |
| H, A, α | Cloude decomposition parameters |
| β, δ, γ | Polarimetric intercorrelation parameters |
| mFP | Barakat degree of polarization |
| θFP | Scattering type parameter based on degree of polarization |
| Ps, Pd, Pv, Pc | Model-free decomposition powers |
| purity | Scattering degree of purity |
| depolarization | Depolarization index |
| Span | Total power |
| HI, HP | Shannon entropy parameters |
This table summarizes the various polarimetric parameters and features used in the study, providing a quick reference for their short form descriptions and corresponding details.
- When asked about a dataloader, perplexity gave it a shot
- Basically makes sense at a simple level