Spaces:

coderuday21
/

satdetect

Sleeping

App Files Files Community

satdetect

Commit History

Jun 15

10:03
Fix water misclassification and reduce detection hallucinations.⁠

coderuday21 +1: Cursor

53d0258
09:54
Restore union-based detection fusion for higher recall.⁠

coderuday21 +1: Cursor

c558d36
09:37
Fix AdaptFormer on production: pin protobuf and expose model health in /health⁠

coderuday21 +1: Cursor

16caeeb

Jun 12

08:12
Align production with no-login app and refresh deploy docs⁠

coderuday21 +1: Cursor

c506ed2
06:44
Remove login; open detection page directly

coderuday21

e55b60b

Jun 10

09:24
Add run.py launcher (IDLE/double-click/terminal entry point)⁠

coderuday21 +1: Cursor

1691e34
05:08
Add WORKFLOW.md: project workflow and onboarding guide⁠

coderuday21 +1: Cursor

7a58fa9

Jun 4

04:17
Fix detection precision: gated fusion, strict registration, calibrated thresholds, preload model⁠

coderuday21 +1: Cursor

ba4abf7

Apr 11

09:03
Fix: rename models/ to cd_models/ to avoid shadowing app/models.py (SQLAlchemy User model)⁠

coderuday21

3e1a5d9
08:52
Add SIFT+FLANN registration, Siamese U-Net, multi-scale detection, ExG vegetation, Hybrid AI mode, confidence maps⁠

coderuday21

aa4d14b
07:00
Improve building detection: fix tile size to 256, lower thresholds, add SSIM, smarter scoring⁠

coderuday21

ce994d8
06:49
Fix: add einops dependency required by AdaptFormer custom code⁠

coderuday21

b7a71f3
06:45
Fix: add trust_remote_code=True for AdaptFormer custom model code⁠

coderuday21

6777e8b
06:41
Production overhaul: pre-trained AdaptFormer model + detection quality improvements⁠

coderuday21

3808a54

Apr 10

11:16
Optimize notebook for CPU training: MobileNetV2 encoder, 15 epochs, batch 4⁠

coderuday21

d555eda
11:01
Fix notebook: use HuggingFace datasets CDN instead of flaky Google Drive links⁠

coderuday21

301b9b6
10:43
Add Siamese U-Net training notebook and model inference integration⁠

coderuday21

1b91ffc
10:34
Clean visualization: subtle tint overlay, outline-only boxes, 30-region cap, high-contrast labels⁠

coderuday21

7d0f836

Apr 8

14:14
Redesign detection pipeline: CVA, vegetation/shadow suppression, SNR fusion, multi-scale SSIM⁠

coderuday21

debf6fe
11:56
Thinner, more precise detection lines and compact labels⁠

coderuday21

8beea23
11:52
Restore Method dropdown, Image Registration and Radiometric Normalization options⁠

coderuday21

dfe265a

Apr 7

19:32
Force cache bust - bump APP_BUILD, CSS and JS versions⁠

coderuday21

8d42f45
19:16
Remove method/model selection UI - use AI-Based Deep Learning by default⁠

coderuday21

0c0a5a2
19:06
Revert to satellite change detection only - remove landslide, pothole, detection type menu⁠

coderuday21

7e52c2b

Apr 1

09:18
Pothole YOLO pipeline: single-image mode, modular detector, red box visualization⁠

coderuday21

466c163
06:16
Add pothole detection as new menu type with separate engine and research plan⁠

coderuday21

4e9c6ba

Mar 30

13:57
HF runtime fix: bind uvicorn to dynamic PORT env⁠

coderuday21

7186eb1
11:55
UX: always land on detection-type menu after login⁠

coderuday21

7496827
11:03
Fix detection-type menu navigation: switch mode and dashboard without reload loop⁠

coderuday21

60a5684
07:44
UX: show detection-type menu after auth then auto-redirect based on URL⁠

coderuday21

236cdf9
07:41
HF: force writable hub cache (/tmp/hf)⁠

coderuday21

064e77c
07:05
HF: relax pip wheel-only; upgrade pip tools⁠

coderuday21

e0949ca
06:59
HF: ensure pip uses binary wheels during build⁠

coderuday21

148396a
06:48
HF: make Docker builds verbose; move APP_BUILD cache-bust before pip⁠

coderuday21

2ea70af
06:44
UX: show detection-type menu after login before redirect⁠

coderuday21

c6682be
06:33
HF: bump APP_BUILD to force rebuild⁠

coderuday21

56c5fcc
06:14
UX: add choose detection type screen and redirect URLs⁠

coderuday21

f207f6c
05:25
HF: bump APP_BUILD to force clean Docker rebuild⁠

coderuday21

4dac521

Mar 27

06:40
Add landslide detection menu, separate engine, and Uttarakhand integration plan⁠

coderuday21

5cee5a6

Mar 26

11:51
AI/Hybrid overhaul: bidirectional AI fusion, stronger hybrid sensitivity, keep large unclassified regions⁠

coderuday21

105a05d
07:07
AI rework: percentile+blur threshold + less aggressive cleanup; Hybrid lower threshold⁠

coderuday21

2d5ff7e
06:46
Fix: fallback to Image Difference when AI/Hybrid mask is too sparse⁠

coderuday21

ba711fa
06:32
Detection fix: ECC registration fallback + less aggressive large-region rejection⁠

coderuday21

18a63c5
06:03
Detection tuning controls + threshold diagnostics in API response⁠

coderuday21

c2dff01

Mar 13

12:36
UI: show pseudo progress percentage while analyzing images⁠

coderuday21

2b175bf
12:32
Email API: attach populated ChangeDetection.html report (base64)⁠

coderuday21

0e5174f

Mar 12

16:09
Email API: send required FileName and AttachmentBase64 placeholders⁠

coderuday21

7392882
16:02
Use manager email API for notifications (multipart/form-data); SMTP fallback⁠

coderuday21

6383573
06:43
Revert Resend; use SMTP only, custom email API to be integrated later⁠

coderuday21

89a1d85
05:29
Use Resend HTTPS API for email (HF Spaces); SMTP fallback for local⁠

coderuday21

2d92257

Commit History

Jun 15

Fix water misclassification and reduce detection hallucinations.⁠

Restore union-based detection fusion for higher recall.⁠

Fix AdaptFormer on production: pin protobuf and expose model health in /health⁠

Jun 12

Align production with no-login app and refresh deploy docs⁠

Remove login; open detection page directly

Jun 10

Add run.py launcher (IDLE/double-click/terminal entry point)⁠

Add WORKFLOW.md: project workflow and onboarding guide⁠

Jun 4

Fix detection precision: gated fusion, strict registration, calibrated thresholds, preload model⁠

Apr 11

Fix: rename models/ to cd_models/ to avoid shadowing app/models.py (SQLAlchemy User model)⁠

Add SIFT+FLANN registration, Siamese U-Net, multi-scale detection, ExG vegetation, Hybrid AI mode, confidence maps⁠

Improve building detection: fix tile size to 256, lower thresholds, add SSIM, smarter scoring⁠

Fix: add einops dependency required by AdaptFormer custom code⁠

Fix: add trust_remote_code=True for AdaptFormer custom model code⁠

Production overhaul: pre-trained AdaptFormer model + detection quality improvements⁠

Apr 10

Optimize notebook for CPU training: MobileNetV2 encoder, 15 epochs, batch 4⁠

Fix notebook: use HuggingFace datasets CDN instead of flaky Google Drive links⁠

Add Siamese U-Net training notebook and model inference integration⁠

Clean visualization: subtle tint overlay, outline-only boxes, 30-region cap, high-contrast labels⁠

Apr 8

Redesign detection pipeline: CVA, vegetation/shadow suppression, SNR fusion, multi-scale SSIM⁠

Thinner, more precise detection lines and compact labels⁠

Restore Method dropdown, Image Registration and Radiometric Normalization options⁠

Apr 7

Force cache bust - bump APP_BUILD, CSS and JS versions⁠

Remove method/model selection UI - use AI-Based Deep Learning by default⁠

Revert to satellite change detection only - remove landslide, pothole, detection type menu⁠

Apr 1

Pothole YOLO pipeline: single-image mode, modular detector, red box visualization⁠

Add pothole detection as new menu type with separate engine and research plan⁠

Mar 30

HF runtime fix: bind uvicorn to dynamic PORT env⁠

UX: always land on detection-type menu after login⁠

Fix detection-type menu navigation: switch mode and dashboard without reload loop⁠

UX: show detection-type menu after auth then auto-redirect based on URL⁠

HF: force writable hub cache (/tmp/hf)⁠

HF: relax pip wheel-only; upgrade pip tools⁠

HF: ensure pip uses binary wheels during build⁠

HF: make Docker builds verbose; move APP_BUILD cache-bust before pip⁠

UX: show detection-type menu after login before redirect⁠

HF: bump APP_BUILD to force rebuild⁠

UX: add choose detection type screen and redirect URLs⁠

HF: bump APP_BUILD to force clean Docker rebuild⁠

Mar 27

Add landslide detection menu, separate engine, and Uttarakhand integration plan⁠

Mar 26

AI/Hybrid overhaul: bidirectional AI fusion, stronger hybrid sensitivity, keep large unclassified regions⁠

AI rework: percentile+blur threshold + less aggressive cleanup; Hybrid lower threshold⁠

Fix: fallback to Image Difference when AI/Hybrid mask is too sparse⁠

Detection fix: ECC registration fallback + less aggressive large-region rejection⁠

Detection tuning controls + threshold diagnostics in API response⁠

Mar 13

UI: show pseudo progress percentage while analyzing images⁠

Email API: attach populated ChangeDetection.html report (base64)⁠

Mar 12

Email API: send required FileName and AttachmentBase64 placeholders⁠

Use manager email API for notifications (multipart/form-data); SMTP fallback⁠

Revert Resend; use SMTP only, custom email API to be integrated later⁠

Use Resend HTTPS API for email (HF Spaces); SMTP fallback for local⁠