# anmolrawat990/sen2sr-spectral-enhancement architecture

> Architecture diagram of the GitHub repository anmolrawat990/sen2sr-spectral-enhancement, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-16). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/anmolrawat990/sen2sr-spectral-enhancement
- Repository: https://github.com/anmolrawat990/sen2sr-spectral-enhancement

## Overview

SEN2SR is a notebook-based remote-sensing research workflow for enhancing Sentinel-2 multispectral imagery, chiefly from 10 m to 2.5 m while preserving radiometry and spectral indices. Users prepare or obtain scenes and benchmark datasets, run pretrained Lite or Full inference, fine-tune with compound spatial/spectral losses, and compare outputs; optional cascaded experiments explore finer resolutions. The graph includes documented capabilities and the sampled WorldStrat extraction utility. Most notebook internals were not sampled, so their links reflect README statements rather than inspected calls.

Main entry points: [`notebooks/01_data_preparation`](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/01_data_preparation) (Scene preparation), [`notebooks/04_finetuning_experiments/Train_SEN2SR_Lite_Compound_Loss.ipynb`](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/04_finetuning_experiments/Train_SEN2SR_Lite_Compound_Loss.ipynb) (Lite fine-tuning), [`notebooks/06_full_model_experiments/Train_SEN2SR_Full_Compound_Loss.ipynb`](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/06_full_model_experiments/Train_SEN2SR_Full_Compound_Loss.ipynb) (Full fine-tuning), [`notebooks/07_cascaded_1.5m_vhr_experiments/Train_SEN2SR_Cascaded_1.5m_VHR.ipynb`](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/07_cascaded_1.5m_vhr_experiments/Train_SEN2SR_Cascaded_1.5m_VHR.ipynb) (1.5m cascaded training)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_data["Data preparation"]
  node_datasets[("Benchmark datasets")]
  node_worldstrat["WorldStrat extractor"]
  node_scene_prep["Scene preparation"]
end

subgraph group_models["Enhancement experiments"]
  node_lite_infer["Lite inference"]
  node_full_infer["Full inference"]
  node_loss_method["Compound loss method"]
  node_cascade_method["Cascaded refinement method"]
  node_lite_train["Lite fine-tuning"]
  node_full_train["Full fine-tuning"]
  node_vhr_train["1.5m cascaded training"]
  node_ultra_train["1m refinement training"]
end

subgraph group_evaluation["Evaluation and results"]
  node_lite_compare["Lite comparison"]
  node_full_compare["Full comparison"]
  node_vhr_compare["1.5m comparison"]
  node_ultra_compare["1m comparison"]
  node_results[("Benchmark results")]
end

node_user(("Researcher"))
node_copernicus[("Copernicus scenes")]

node_user -->|"downloads scenes"| node_copernicus
node_user -->|"prepares data"| node_scene_prep
node_copernicus -->|"supplies imagery"| node_scene_prep
node_scene_prep -->|"prepares datasets"| node_datasets
node_worldstrat -->|"extracts archive"| node_datasets
node_user -->|"runs inference"| node_lite_infer
node_user -->|"runs inference"| node_full_infer
node_user -->|"fine-tunes model"| node_lite_train
node_loss_method -->|"guides training"| node_lite_train
node_user -->|"fine-tunes model"| node_full_train
node_loss_method -->|"guides training"| node_full_train
node_cascade_method -->|"guides refinement"| node_vhr_train
node_user -->|"runs experiment"| node_vhr_train
node_user -->|"runs experiment"| node_ultra_train
node_lite_train -->|"compares fine-tuned model"| node_lite_compare
node_lite_infer -->|"provides baseline"| node_lite_compare
node_full_train -->|"compares fine-tuned model"| node_full_compare
node_full_infer -->|"provides baseline"| node_full_compare
node_vhr_train -->|"evaluates stages"| node_vhr_compare
node_ultra_train -->|"evaluates refinement"| node_ultra_compare
node_lite_compare -.->|"records metrics"| node_results
node_full_compare -->|"records benchmarks"| node_results

click node_datasets "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/01_data_preparation"
click node_worldstrat "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/scripts/extract_worldstrat.py"
click node_scene_prep "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/01_data_preparation"
click node_lite_infer "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/02_baseline_inference"
click node_full_infer "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/02_baseline_inference"
click node_loss_method "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/03_methodology_docs"
click node_cascade_method "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/03_methodology_docs"
click node_lite_train "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/04_finetuning_experiments/Train_SEN2SR_Lite_Compound_Loss.ipynb"
click node_full_train "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/06_full_model_experiments/Train_SEN2SR_Full_Compound_Loss.ipynb"
click node_vhr_train "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/07_cascaded_1.5m_vhr_experiments/Train_SEN2SR_Cascaded_1.5m_VHR.ipynb"
click node_ultra_train "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/05_ultra_highres_1m_experiments/Train_1Meter_Cascaded_Refiner.ipynb"
click node_lite_compare "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/04_finetuning_experiments/SEN2SR_Baseline_vs_Finetuned_Comparison.ipynb"
click node_full_compare "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/06_full_model_experiments/SEN2SR_Full_Baseline_vs_Finetuned_Comparison.ipynb"
click node_vhr_compare "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/07_cascaded_1.5m_vhr_experiments/SEN2SR_1.5Meter_MultiStage_Comparison.ipynb"
click node_ultra_compare "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/05_ultra_highres_1m_experiments/SEN2SR_1Meter_4Way_Comparison.ipynb"
click node_results "https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/results"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_datasets,node_worldstrat,node_scene_prep toneBlue
class node_lite_infer,node_full_infer,node_loss_method,node_cascade_method,node_lite_train,node_full_train,node_vhr_train,node_ultra_train,node_copernicus toneAmber
class node_lite_compare,node_full_compare,node_vhr_compare,node_ultra_compare,node_results toneMint
class node_user toneIndigo
```

## Components

### Data preparation

- [Benchmark datasets](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/01_data_preparation): `notebooks/01_data_preparation`
- [WorldStrat extractor](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/scripts/extract_worldstrat.py): `scripts/extract_worldstrat.py`
- [Scene preparation](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/01_data_preparation): `notebooks/01_data_preparation`

### Enhancement experiments

- [Lite inference](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/02_baseline_inference): `notebooks/02_baseline_inference`
- [Full inference](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/02_baseline_inference): `notebooks/02_baseline_inference`
- [Compound loss method](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/03_methodology_docs): `notebooks/03_methodology_docs`
- [Cascaded refinement method](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/notebooks/03_methodology_docs): `notebooks/03_methodology_docs`
- [Lite fine-tuning](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/04_finetuning_experiments/Train_SEN2SR_Lite_Compound_Loss.ipynb): `notebooks/04_finetuning_experiments/Train_SEN2SR_Lite_Compound_Loss.ipynb`
- [Full fine-tuning](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/06_full_model_experiments/Train_SEN2SR_Full_Compound_Loss.ipynb): `notebooks/06_full_model_experiments/Train_SEN2SR_Full_Compound_Loss.ipynb`
- [1.5m cascaded training](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/07_cascaded_1.5m_vhr_experiments/Train_SEN2SR_Cascaded_1.5m_VHR.ipynb): `notebooks/07_cascaded_1.5m_vhr_experiments/Train_SEN2SR_Cascaded_1.5m_VHR.ipynb`
- [1m refinement training](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/05_ultra_highres_1m_experiments/Train_1Meter_Cascaded_Refiner.ipynb): `notebooks/05_ultra_highres_1m_experiments/Train_1Meter_Cascaded_Refiner.ipynb`

### Evaluation and results

- [Lite comparison](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/04_finetuning_experiments/SEN2SR_Baseline_vs_Finetuned_Comparison.ipynb): `notebooks/04_finetuning_experiments/SEN2SR_Baseline_vs_Finetuned_Comparison.ipynb`
- [Full comparison](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/06_full_model_experiments/SEN2SR_Full_Baseline_vs_Finetuned_Comparison.ipynb): `notebooks/06_full_model_experiments/SEN2SR_Full_Baseline_vs_Finetuned_Comparison.ipynb`
- [1.5m comparison](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/07_cascaded_1.5m_vhr_experiments/SEN2SR_1.5Meter_MultiStage_Comparison.ipynb): `notebooks/07_cascaded_1.5m_vhr_experiments/SEN2SR_1.5Meter_MultiStage_Comparison.ipynb`
- [1m comparison](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/blob/main/notebooks/05_ultra_highres_1m_experiments/SEN2SR_1Meter_4Way_Comparison.ipynb): `notebooks/05_ultra_highres_1m_experiments/SEN2SR_1Meter_4Way_Comparison.ipynb`
- [Benchmark results](https://github.com/anmolrawat990/sen2sr-spectral-enhancement/tree/main/results): `results`

### Other components

- **Researcher**
- **Copernicus scenes**

## Connections

- Researcher → Copernicus scenes: downloads scenes (evidence: `README.md`)
- Researcher → Scene preparation: prepares data (evidence: `README.md`)
- Copernicus scenes → Scene preparation: supplies imagery (evidence: `README.md`)
- Scene preparation → Benchmark datasets: prepares datasets (evidence: `README.md`)
- WorldStrat extractor → Benchmark datasets: extracts archive (evidence: `scripts/extract_worldstrat.py`)
- Researcher → Lite inference: runs inference (evidence: `README.md`)
- Researcher → Full inference: runs inference (evidence: `README.md`)
- Researcher → Lite fine-tuning: fine-tunes model (evidence: `README.md`)
- Compound loss method → Lite fine-tuning: guides training (evidence: `README.md`)
- Researcher → Full fine-tuning: fine-tunes model (evidence: `README.md`)
- Compound loss method → Full fine-tuning: guides training (evidence: `README.md`)
- Cascaded refinement method → 1.5m cascaded training: guides refinement (evidence: `README.md`)
- Researcher → 1.5m cascaded training: runs experiment (evidence: `README.md`)
- Researcher → 1m refinement training: runs experiment (evidence: `README.md`)
- Lite fine-tuning → Lite comparison: compares fine-tuned model (evidence: `README.md`)
- Lite inference → Lite comparison: provides baseline (evidence: `README.md`)
- Full fine-tuning → Full comparison: compares fine-tuned model (evidence: `README.md`)
- Full inference → Full comparison: provides baseline (evidence: `README.md`)
- 1.5m cascaded training → 1.5m comparison: evaluates stages (evidence: `README.md`)
- 1m refinement training → 1m comparison: evaluates refinement (evidence: `README.md`)
- Lite comparison → Benchmark results: records metrics
- Full comparison → Benchmark results: records benchmarks (evidence: `README.md`)
