Multi-Modal Foundation · isp-uv-es
StarCOP
A segmentation model for detecting methane plumes in airborne and satellite hyperspectral remote sensing imagery.
What StarCOP is
StarCOP (specifically the HyperSTARCOP variants) is a machine learning model developed for the automated semantic segmentation of methane (CH4) plumes. Created by researchers from isp-uv-es and collaborating institutions, it identifies point-source greenhouse gas emissions captured by hyperspectral instruments.
Automating the detection of methane emissions is essential for climate research and environmental monitoring, as point-source releases contribute significantly to atmospheric greenhouse gas concentrations. StarCOP provides an automated approach to segmenting plume boundaries from airborne sensors such as AVIRIS and spaceborne instruments such as EMIT, including workflows that operate directly on raw NetCDF data files.
How it works
The model performs semantic segmentation on hyperspectral data inputs to isolate methane plumes at pixel level. Available variants include configurations trained solely on mag1c data (HyperSTARCOP, only mag1c) and configurations incorporating both mag1c and RGB bands (HyperSTARCOP, mag1c + rgb).
Evaluated using test datasets from both the AVIRIS and EMIT instruments, the system processes hyperspectral observations to delineate greenhouse gas plumes across remote sensing scenes.
Key facts
- Task
- Semantic image segmentation
- Target gas
- Methane (CH4)
- Supported instruments
- EMIT and AVIRIS
- Model variants
- HyperSTARCOP (only mag1c) and HyperSTARCOP (mag1c + rgb)
- Release year
- 2023
What you can use it for
- Spaceborne methane plume detection
- Segment point-source methane emissions in space-based hyperspectral data from the EMIT sensor.
- Airborne emission monitoring
- Identify and map methane plumes in high-resolution airborne imagery collected by AVIRIS flights.
- Direct NetCDF processing
- Perform automated inference for methane plume identification directly on raw EMIT NetCDF (.nc) files.
Run it here
We list this model but cannot run it yet: no library and no architectures — nothing can load it.
Everything on this page still applies. Running it needs a loader for this repository, not a bigger machine.
Strengths and limits
Good at
- Supports multiple input modalities, offering versions for mag1c data alone as well as combined mag1c and RGB imagery.
- Demonstrated applicability across both airborne (AVIRIS) and satellite (EMIT) hyperspectral platforms.
- Enables automated pixel-level semantic segmentation of point-source greenhouse gas emissions.
Watch out for
- Restricted to hyperspectral platforms specifically matching AVIRIS and EMIT data characteristics.
- Requires specific preprocessing inputs such as mag1c or paired mag1c and RGB bands.
Inputs and outputs
You provide
Image
PNG, JPEG or TIFF. Sent to the model as-is.
You get back
The returned image.
Labels with confidence scores, highest first.
Licence
Check the licence before commercial use · cc-by-nc-4.0
The cc-by-nc-4.0 licence carries terms worth reading before you depend on it.
Questions
Can I use StarCOP commercially?
Check the licence before commercial use. The cc-by-nc-4.0 licence carries terms worth reading before you depend on it.
Which instruments are supported by StarCOP?
StarCOP is trained and evaluated on hyperspectral data from the airborne AVIRIS instrument and the spaceborne EMIT sensor.
What model variants are available?
The repository includes two HyperSTARCOP variants: one trained solely on mag1c data, and another trained on mag1c combined with RGB imagery.
Can StarCOP run directly on raw sensor files?
Yes, inference workflows are provided to run StarCOP models directly on raw EMIT NetCDF (.nc) data files.
How to cite it
Růžička, V., Mateo-Garcia, G., Gómez-Chova, L., Vaughan, A., Guanter, L., & Markham, A. (2023). Semantic segmentation of methane plumes with hyperspectral machine learning models. Scientific Reports, 13, 19999. https://doi.org/10.1038/s41598-023-44918-6
Written from StarCOP's own documentation (model card). Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.