Result images
Browse files- .gitattributes +6 -0
- README.md +21 -0
- images/1_city.jpg +3 -0
- images/2_slum_and_industry.jpg +3 -0
- images/3_stadium.jpg +3 -0
- images/4_village.jpg +3 -0
- images/dharavi_live.png +3 -0
- images/vs_other_methods.jpg +3 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
images/vs_other_methods.jpg filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
images/1_city.jpg filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
images/2_slum_and_industry.jpg filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
images/3_stadium.jpg filter=lfs diff=lfs merge=lfs -text
|
| 40 |
+
images/4_village.jpg filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
images/dharavi_live.png filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
|
@@ -19,6 +19,11 @@ rewrites what the evidence contradicts, and outputs a per-pixel expected error (
|
|
| 19 |
|
| 20 |
Code, training and inference: **https://github.com/Harp404/sentinel-superres**
|
| 21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
## Files (everything the pipeline needs, in one place)
|
| 23 |
|
| 24 |
| Component | File | Original | Licence |
|
|
@@ -49,6 +54,22 @@ use for military, warfare or espionage purposes and requires acknowledging DINOv
|
|
| 49 |
| SUPIR alone | 15.56% | 47.9% |
|
| 50 |
| **SentinelSharp** | **12.07%** | **54.5%** |
|
| 51 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
## Load
|
| 53 |
|
| 54 |
```python
|
|
|
|
| 19 |
|
| 20 |
Code, training and inference: **https://github.com/Harp404/sentinel-superres**
|
| 21 |
|
| 22 |
+

|
| 23 |
+
|
| 24 |
+
*Same Sentinel-2 input, five methods, five held-out places. Satlas and SUPIR look sharp but invent buildings; ESA's
|
| 25 |
+
SEN2SR stays blurry; SentinelSharp is sharp and matches the real 2.5 m SPOT photo.*
|
| 26 |
+
|
| 27 |
## Files (everything the pipeline needs, in one place)
|
| 28 |
|
| 29 |
| Component | File | Original | Licence |
|
|
|
|
| 54 |
| SUPIR alone | 15.56% | 47.9% |
|
| 55 |
| **SentinelSharp** | **12.07%** | **54.5%** |
|
| 56 |
|
| 57 |
+
Left: Sentinel-2 input (10 m). Middle: SentinelSharp (2.5 m). Right: SPOT 6/7 ground truth (2.5 m), photographed within
|
| 58 |
+
days of the Sentinel-2 passes.
|
| 59 |
+
|
| 60 |
+

|
| 61 |
+

|
| 62 |
+

|
| 63 |
+

|
| 64 |
+
|
| 65 |
+
Live run on Dharavi (Mumbai): SUPIR alone invents a car park; the verification head restores the informal housing.
|
| 66 |
+
|
| 67 |
+

|
| 68 |
+
|
| 69 |
+
*Images: contains modified Copernicus Sentinel data (2019–2026). SPOT 6/7 imagery © Airbus DS, via WorldStrat
|
| 70 |
+
(CC BY-NC 4.0). Dharavi reference photo: Esri World Imagery (Esri, Maxar, Earthstar Geographics).*
|
| 71 |
+
|
| 72 |
+
|
| 73 |
## Load
|
| 74 |
|
| 75 |
```python
|
images/1_city.jpg
ADDED
|
Git LFS Details
|
images/2_slum_and_industry.jpg
ADDED
|
Git LFS Details
|
images/3_stadium.jpg
ADDED
|
Git LFS Details
|
images/4_village.jpg
ADDED
|
Git LFS Details
|
images/dharavi_live.png
ADDED
|
Git LFS Details
|
images/vs_other_methods.jpg
ADDED
|
Git LFS Details
|